Source code for evalml.pipelines.components.estimators.classifiers.logistic_regression

import numpy as np
from sklearn.linear_model import LogisticRegression as LogisticRegression
from skopt.space import Real

from evalml.model_types import ModelTypes
from evalml.pipelines.components.estimators import Estimator
from evalml.problem_types import ProblemTypes


class LogisticRegressionClassifier(Estimator):
    """
    Logistic Regression Classifier
    """
    name = "Logistic Regression Classifier"
    hyperparameter_ranges = {
        "penalty": ["l2"],
        "C": Real(.01, 10),
    }
    model_type = ModelTypes.LINEAR_MODEL
    problem_types = [ProblemTypes.BINARY, ProblemTypes.MULTICLASS]

    def __init__(self, penalty="l2", C=1.0, n_jobs=-1, random_state=0):
        parameters = {"penalty": penalty,
                      "C": C}
        lr_classifier = LogisticRegression(penalty=penalty,
                                           C=C,
                                           random_state=random_state,
                                           multi_class="auto",
                                           solver="lbfgs",
                                           n_jobs=n_jobs)
        super().__init__(parameters=parameters,
                         component_obj=lr_classifier,
                         random_state=random_state)

    @property
    def feature_importances(self):
        coef_ = self._component_obj.coef_
        # binary classification case
        if len(coef_) <= 2:
            return coef_[0]
        else:
            # mutliclass classification case
            return np.linalg.norm(coef_, axis=0, ord=2)