Logistic Regression.py

from sklearn.linear_model import LogisticRegression

class LR(LogisticRegression):
    def __init__(self, threshold=0.01, dual=False, tol=1e-4, C=1.0,
                 fit_intercept=True, intercept_scaling=1, class_weight=None,
                 random_state=None, solver='liblinear', max_iter=100,
                 multi_class='ovr', verbose=0, warm_start=False, n_jobs=1):

        #Thresold
        self.threshold = threshold
        LogisticRegression.__init__(self, penalty='l1', dual=dual, tol=tol, C=C,
                 fit_intercept=fit_intercept, intercept_scaling=intercept_scaling, class_weight=class_weight,
                 random_state=random_state, solver=solver, max_iter=max_iter,
                 multi_class=multi_class, verbose=verbose, warm_start=warm_start, n_jobs=n_jobs)
        #Create L2 logistic regression using the same parameters
        self.l2 = LogisticRegression(penalty='l2', dual=dual, tol=tol, C=C, fit_intercept=fit_intercept, intercept_scaling=intercept_scaling, class_weight = class_weight, random_state=random_state, solver=solver, max_iter=max_iter, multi_class=multi_class, verbose=verbose, warm_start=warm_start, n_jobs=n_jobs)

    def fit(self, X, y, sample_weight=None):
        # Training L1 logistic regression
        super(LR, self).fit(X, y, sample_weight=sample_weight)
        self.coef_old_ = self.coef_.copy()
        # L2 logistic regression training
        self.l2.fit(X, y, sample_weight=sample_weight)

        cntOfRow, cntOfCol = self.coef_.shape
        #Number of coefficient matrix The number of rows corresponds to the number of types of target values
        for i in range(cntOfRow):
            for j in range(cntOfCol):
                coef = self.coef_[i][j]
                # The weight coefficient of L1 logistic regression is not 0.
                if coef != 0:
                    idx = [j]
                    #correspond to the weight coefficient in L2 logistic regression
                    coef1 = self.l2.coef_[i][j]
                    for k in range(cntOfCol):
                        coef2 = self.l2.coef_[i][k]
                        #In L2 logistic regression, the difference between the weight coefficients is less than the set threshold, and the corresponding weight in L1 is 0.
                        if abs(coef1-coef2) < self.threshold and j != k and self.coef_[i][k] == 0:
                            idx.append(k)
                    #Calculate the mean value of the weight coefficient of this type of feature
                    mean = coef / len(idx)
                    self.coef_[i][idx] = mean
        return self