classification.py 文件源码

def k_fold_classification(x, y, folds, classifier_name='logistic_regression', bootstrap=False):
    x_train_list, y_train_list, x_test_list, y_test_list = k_fold_sample_data_set(x, y, folds)
    model_performance_dict = dict()
    total_accuracy = 0
    for j in range(0, folds, 1):
        # split data set in train and test set
        if bootstrap:
            x_train, y_train, x_test, y_test = random_sample_data_set(x, y, folds)
        else:
            x_train = x_train_list[j]
            y_train = y_train_list[j]
            x_test = x_test_list[j]
            y_test = y_test_list[j]
        x_train, x_test = scale_sets(x_train, x_test, classifier_name)
        model = model_fitting(x_train, y_train, classifier_name)
        predicted_labels = model.predict(x_test)
        print(metrics.accuracy_score(y_test, predicted_labels))
        total_accuracy += metrics.accuracy_score(y_test, predicted_labels)
    model_performance_dict["accuracy"] = float(total_accuracy)/float(folds)
    export_model_performance(model_performance_dict)