marketing_predict.py 文件源码

def learn_decision_tree(data):
    DT = tree.DecisionTreeClassifier(max_depth=7)
    scorer = make_scorer(matthews_corrcoef)
    for i in range(5):
        scores = cross_val_score(DT, data.X_train, data.y_train, cv=10, scoring=scorer)
        print("iteration",i, "dt mean:", scores.mean())
        scores = list(scores)
        print("Decision Tree train scores:\n", scores)
    return DT
    # DT = DT.fit(train_data[:, :-1], train_data[:, -1])
    # predictionsDT = DT.predict(validation_data[:, :-1])

    # validating predicions
    # dtError = 0
    # for i in range(0, len(validation_data)):
    #         if(validation_data[i][20] != predictionsDT[i]):
    #                 dtError = dtError + 1
    # print("DT Error : ", float(dtError)/len(validation_data)*100.0)