gbc.py 文件源码

def GradientBoostingClassifier(X_train, y_train, X_test):
    from sklearn.ensemble import GradientBoostingClassifier
    now = datetime.datetime.now()
    print ("GradientBoostingClassifier start in " + now.strftime('%Y-%m-%d %H:%M:%S'))
    GBC = GradientBoostingClassifier(max_features = 'sqrt',
                                     n_estimators = 300,
                                     learning_rate = 0.02,
                                     max_depth = 8,
                                     subsample = 0.8,
                                     n_jobs =4)
    GBC.fit(X_train, y_train)
    now = datetime.datetime.now()
    print ("GradientBoostingClassifier train done in " + now.strftime('%Y-%m-%d %H:%M:%S'))

    y_pred_GBC = GBC.predict_proba(X_test)
    y_pred_GBC = pd.DataFrame(y_pred_GBC[:,1:2],columns=['GBC_predictions'])
    y_pred_GBC.to_csv('GBC_result_all.csv', index=False)
    now = datetime.datetime.now()
    print ("GradientBoostingClassifier predict done in " + now.strftime('%Y-%m-%d %H:%M:%S'))