Cross-validation

*PathCV estimators wrap a path solver in K-fold cross-validation. Two flavors of selection:

• Lower-is-better metrics (MSE, binomial deviance, Poisson deviance): pick λ minimizing mean test score.

• Higher-is-better metrics (Cox concordance / c-index): pick λ maximizing.

The metric is auto-selected by family. After fitting, a final refit on the full data at λ_best is stored on coef_ / intercept_.

36 estimators total — across the LS, logistic, Poisson, Cox, and multi-task families. They all share the same fit/predict surface as their non-CV counterparts.

Threaded fold parallelism. Every CV class accepts an n_jobs parameter (default None = serial). The fold loop dispatches via joblib.Parallel(prefer="threads"), and the underlying Rust path solvers release the GIL during compute (py.allow_threads in every PyO3 entry point), so threads actually run concurrently. Bitwise parity between n_jobs=1 and n_jobs=-1 is verified in the test suite. Typical 5-fold speedup is ~2.3–2.5× on 8 cores, bounded by the number of folds and the inner-path’s rayon parallelism already saturating cores per fit.

LS family

class skein_glm.cv.MCPPathCV(gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold cross-validated MCP path. Picks the λ minimizing mean test MSE and exposes the corresponding β as coef_ / intercept_.

The underlying solver is MCPPathRegressor. All of its constructor knobs are forwarded; CV adds cv (int K or sklearn splitter) and random_state (passed to KFold shuffle when cv is an int).

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.SCADPathCV(a=3.7, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold cross-validated SCAD path. Same shape as MCPPathCV, but the underlying solver is SCADPathRegressor.

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (SCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (SCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (SCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.ElasticNetPathCV(alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold cross-validated elastic-net path. Same shape as MCPPathCV, but the underlying solver is ElasticNetPathRegressor. Picks the λ minimizing mean test MSE on the supplied folds.

Parameters:

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (ElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (ElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (ElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.BridgePathCV(q=0.5, *, eps=1e-06, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=10, outer_tol=1e-06, fit_intercept=True, standardize=False, acceleration=5)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold cross-validated bridge (ℓ_q) path. Picks the λ minimizing mean test MSE on the supplied folds. The non-convex inner objective means the chosen λ is a local-min selection — initialization (warm-starting from large λ down) makes this stable in practice but not guaranteed to be the global optimum at any λ.

Parameters:

• q (float)

• eps (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• acceleration (int | None)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (BridgePathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (BridgePathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (BridgePathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.GroupLassoPathCV(groups, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a group lasso λ-path. Picks the λ minimizing mean test MSE.

Parameters:

• groups (NDArray[np.int64])

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (GroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (GroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (GroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.GroupMCPPathCV(groups, gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False, max_outer=10, outer_tol=1e-06)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a group MCP λ-path. The inner path solver is the native group-MCP block-CD shipped in v0.8 (no LLA outer loop) — see skein_glm.GroupMCPPathRegressor for the algorithmic details and the max_outer / outer_tol backward-compat notes.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

• max_outer (int)

• outer_tol (float)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (GroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (GroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (GroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.GroupSCADPathCV(groups, a=3.7, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False, max_outer=10, outer_tol=1e-06)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a group SCAD λ-path (LLA outer loop). SCAD shape a > 2 (default 3.7).

Parameters:

• groups (NDArray[np.int64])

• a (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

• max_outer (int)

• outer_tol (float)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (GroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (GroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (GroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.GroupElasticNetPathCV(groups, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a group elastic-net λ-path. Picks the λ minimizing mean test MSE.

Parameters:

• groups (NDArray[np.int64])

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (GroupElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (GroupElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (GroupElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.SparseGroupLassoPathCV(groups, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a sparse-group lasso λ-path.

Parameters:

• groups (NDArray[np.int64])

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (SparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (SparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (SparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.SparseGroupMCPPathCV(groups, gamma=3.0, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False, max_outer=10, outer_tol=1e-06)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a sparse-group MCP λ-path (LLA outer loop).

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

• max_outer (int)

• outer_tol (float)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (SparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (SparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (SparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.SparseGroupSCADPathCV(groups, a=3.7, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False, max_outer=10, outer_tol=1e-06)

Bases: _PathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a sparse-group SCAD λ-path (LLA outer loop). SCAD shape a > 2 (default 3.7).

Parameters:

• groups (NDArray[np.int64])

• a (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

• max_outer (int)

• outer_tol (float)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (SparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (SparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (SparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

Multi-task LS family

K-fold CV scoring by mean MSE averaged across tasks. Multi-task estimators take 2D Y ∈ ℝ^(n, K) and produce coef_ of shape (K, p). See the multi-task concept page for the data shape and convention notes.

class skein_glm.multitask.MultiTaskLassoPathCV(*, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _MultiTaskPathCVBase

K-fold CV over a multi-task lasso λ-path. Picks the λ minimizing mean test MSE (averaged across tasks).

Parameters:

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultiTaskLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultiTaskLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultiTaskLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multitask.MultiTaskMCPPathCV(gamma=3.0, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=10, outer_tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _MultiTaskPathCVBase

K-fold CV over a multi-task MCP λ-path (LLA outer loop).

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultiTaskMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultiTaskMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultiTaskMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multitask.MultiTaskSCADPathCV(a=3.7, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=10, outer_tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _MultiTaskPathCVBase

K-fold CV over a multi-task SCAD λ-path (LLA outer loop).

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultiTaskSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultiTaskSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultiTaskSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multitask.MultiTaskElasticNetPathCV(alpha=0.5, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, standardize=False, screening='strong', acceleration=5, parallel=False)

Bases: _MultiTaskPathCVBase

K-fold CV over a multi-task elastic-net λ-path.

Parameters:

• alpha (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• standardize (bool)

• screening (str)

• acceleration (int | None)

• parallel (bool)

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultiTaskElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultiTaskElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultiTaskElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

Logistic family

class skein_glm.cv.LogisticMCPPathCV(gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic-MCP path. Picks the λ minimizing mean test binomial deviance.

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticSCADPathCV(a=3.7, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic-SCAD path.

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticElasticNetPathCV(alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + elastic-net path.

Parameters:

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticLassoPathCV(*, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic-lasso path (proper convex L1, not the MCP-at-large-γ approximation).

Parameters:

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticGroupLassoPathCV(groups, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticGroupMCPPathCV(groups, gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + group-MCP path (LLA outer loop).

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticSparseGroupLassoPathCV(groups, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + sparse-group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticSparseGroupMCPPathCV(groups, gamma=3.0, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + sparse-group-MCP path.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.LogisticSparseGroupSCADPathCV(groups, a=3.7, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _LogisticPathCVMixin, BaseEstimator, ClassifierMixin

K-fold CV over a logistic + sparse-group-SCAD path.

Parameters:

• groups (NDArray[np.int64])

• a (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (LogisticSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (LogisticSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (LogisticSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (LogisticSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (LogisticSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

Poisson family

class skein_glm.cv.PoissonMCPPathCV(gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson-MCP path. Picks λ minimizing mean test Poisson deviance.

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonSCADPathCV(a=3.7, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson-SCAD path.

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonElasticNetPathCV(alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + elastic-net path.

Parameters:

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonLassoPathCV(*, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson-lasso path (proper convex L1).

Parameters:

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonGroupLassoPathCV(groups, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonGroupMCPPathCV(groups, gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + group-MCP path.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonSparseGroupLassoPathCV(groups, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + sparse-group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonSparseGroupMCPPathCV(groups, gamma=3.0, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + sparse-group-MCP path.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.PoissonSparseGroupSCADPathCV(groups, a=3.7, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, offset=None, weights=None, coord_weights=None, max_iter=100, tol=1e-06, fit_intercept=True, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _PoissonPathCVMixin, BaseEstimator, RegressorMixin

K-fold CV over a Poisson + sparse-group-SCAD path.

Parameters:

• groups (NDArray[np.int64])

• a (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• offset (NDArray[np.float64] | None)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• fit_intercept (bool)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (PoissonSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (PoissonSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (PoissonSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (PoissonSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

Multinomial family

K-fold CV scoring by mean multinomial deviance. Default splitter is StratifiedKFold so heavy class imbalance doesn’t produce class-empty train folds; folds with fewer than K distinct training classes are defensively skipped (their per-λ scores become NaN and np.nanmean aggregation handles it). See the multinomial concept page for the symmetric softmax parameterization.

class skein_glm.multinomial.MultinomialLassoPathCV(*, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=25, outer_tol=1e-06, fit_intercept=True, standardize=False, acceleration=5)

Bases: _MultinomialPathCVBase

K-fold CV over a multinomial-lasso λ-path, scored by multinomial deviance.

Parameters:

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• acceleration (int | None)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (MultinomialLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultinomialLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (MultinomialLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultinomialLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultinomialLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multinomial.MultinomialMCPPathCV(gamma=3.0, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=25, outer_tol=1e-06, fit_intercept=True, standardize=False, acceleration=5)

Bases: _MultinomialPathCVBase

K-fold CV over a multinomial-MCP λ-path.

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• acceleration (int | None)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (MultinomialMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultinomialMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (MultinomialMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultinomialMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultinomialMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multinomial.MultinomialSCADPathCV(a=3.7, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=25, outer_tol=1e-06, fit_intercept=True, standardize=False, acceleration=5)

Bases: _MultinomialPathCVBase

K-fold CV over a multinomial-SCAD λ-path.

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• acceleration (int | None)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (MultinomialSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultinomialSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (MultinomialSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultinomialSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultinomialSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.multinomial.MultinomialElasticNetPathCV(alpha=0.5, *, cv=5, random_state=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, weights=None, max_iter=100, tol=1e-06, max_outer=25, outer_tol=1e-06, fit_intercept=True, standardize=False, acceleration=5)

Bases: _MultinomialPathCVBase

K-fold CV over a multinomial elastic-net λ-path.

Parameters:

• alpha (float)

• cv (Any)

• random_state (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• max_outer (int)

• outer_tol (float)

• fit_intercept (bool)

• standardize (bool)

• acceleration (int | None)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (MultinomialElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (MultinomialElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_proba_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict_proba method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict_proba if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict_proba.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict_proba.

• self (MultinomialElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (MultinomialElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

set_score_request(*, sample_weight='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

• self (MultinomialElasticNetPathCV)

Returns:

self – The updated object.

Return type:

object

Cox family

Cox CV uses StratifiedKFold by event indicator (so heavy censoring doesn’t produce event-empty train folds). Folds with zero events are defensively skipped.

class skein_glm.cv.CoxMCPPathCV(gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox-MCP path. Picks λ maximizing mean test Harrell concordance index.

Parameters:

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxSCADPathCV(a=3.7, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox-SCAD path.

Parameters:

• a (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxSCADPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxGroupLassoPathCV(groups, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox + group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxGroupMCPPathCV(groups, gamma=3.0, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox + group-MCP path.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxSparseGroupLassoPathCV(groups, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox + sparse-group-lasso path.

Parameters:

• groups (NDArray[np.int64])

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxSparseGroupLassoPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxSparseGroupMCPPathCV(groups, gamma=3.0, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, coord_weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox + sparse-group-MCP path.

Parameters:

• groups (NDArray[np.int64])

• gamma (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxSparseGroupMCPPathCV)

Returns:

self – The updated object.

Return type:

object

class skein_glm.cv.CoxSparseGroupSCADPathCV(groups, a=3.7, alpha=0.5, *, cv=5, random_state=None, n_jobs=None, lambdas=None, n_lambdas=100, lambda_min_ratio=0.001, ties='breslow', weights=None, coord_weights=None, max_iter=100, tol=1e-06, acceleration=5, max_outer=10, outer_tol=1e-06)

Bases: _CoxPathCVMixin, BaseEstimator

K-fold CV over a Cox + sparse-group-SCAD path.

Parameters:

• groups (NDArray[np.int64])

• a (float)

• alpha (float)

• cv (Any)

• random_state (int | None)

• n_jobs (int | None)

• lambdas (NDArray[np.float64] | None)

• n_lambdas (int)

• lambda_min_ratio (float)

• ties (str)

• weights (NDArray[np.float64] | None)

• coord_weights (NDArray[np.float64] | None)

• max_iter (int)

• tol (float)

• acceleration (int | None)

• max_outer (int)

• outer_tol (float)

set_decision_function_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the decision_function method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to decision_function if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to decision_function.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in decision_function.

• self (CoxSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_fit_request(*, event='$UNCHANGED$', time='$UNCHANGED$', x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to fit.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• event (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for event parameter in fit.

• time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in fit.

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in fit.

• self (CoxSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object

set_predict_request(*, x='$UNCHANGED$')

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

• x (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for x parameter in predict.

• self (CoxSparseGroupSCADPathCV)

Returns:

self – The updated object.

Return type:

object