SuperGLMRegressor#

Bases: BaseEstimator, RegressorMixin

Penalised GLM for count / continuous regression.

Accepts DataFrame or ndarray input. See module docstring for details on ndarray mode and penalty defaults.

Parameters:

familystr or Distribution: Distribution family. Strings: "poisson", "gamma", "gaussian". For parameterized families use objects: families.tweedie(p=1.5), families.nb2(theta=1.0). For binary classification use SuperGLMClassifier.
penaltystr, Penalty, or None: Penalty type. None (default) means no feature-selection penalty when selection_penalty is also None or zero. Explicit automatic or positive selection strength auto-upgrades to "group_lasso".
selection_penaltyfloat, {“auto”}, or None: Group penalty strength. None and zero disable feature selection; "auto" explicitly calibrates the strength from the fit data.
spline_penaltyfloat or None: Within-group spline smoothing. Defaults to 0.1.
featuresdict[str, FeatureSpec] or None: Native-style feature specs. Mutually exclusive with the shorthand wrapper arguments (spline_features, categorical_features, numeric_features, non-default n_knots/degree/categorical_base).
spline_featureslist of str or int, or None: Columns to treat as spline features (by name or index).
categorical_featureslist of str or int, or None: Columns to treat as categorical (by name or index). Required for ndarray input if any columns are categorical.
numeric_featureslist of str or int, or None: Columns to treat as numeric (by name or index). Unspecified columns default to numeric.
feature_nameslist of str or None: Column names for ndarray input. Ignored for native DataFrames.
n_knotsint or list of int: Interior knots for spline features.
degreeint: B-spline degree (default 3).
categorical_basestr: Base level strategy ("most_exposed" or "first").
offsetstr, int, list, or None: Offset column(s) by name or index.
weight_semantics{“prior”, “frequency”}: What sample_weight says about a row. "prior" (default) reads it as a precision, Var(Y_i) = phi V(mu_i) / w_i; "frequency" reads it as a replication count, where an integer weight is exactly a repeated row. Passed through to SuperGLM.

summary( alpha: float = 0.05, detail: str = 'compact', level_display: str = 'expanded', )#: Return the fitted model summary with categorical display control.

set_fit_request( *, sample_weight: bool | None | str = '$UNCHANGED$', ) → SuperGLMRegressor#

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:

sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for sample_weight parameter in fit.

Returns:

selfobject: The updated object.

set_score_request( *, sample_weight: bool | None | str = '$UNCHANGED$', ) → SuperGLMRegressor#

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_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for sample_weight parameter in score.

Returns:

selfobject: The updated object.