machine_learning.py 文件源码

def random_forests(wine_set):
    # recode quality (response variable) into 2 groups: 0:{3,4,5}, 1:{6,7,8,9}
    recode = {3: 0, 4: 0, 5: 0, 6: 1, 7: 1, 8: 1, 9: 1}
    wine_set['quality_c'] = wine_set['quality'].map(recode)

    # split into training and testing sets
    predictors = wine_set[["density", 'alcohol', 'sulphates', 'pH', 'volatile_acidity', 'chlorides', 'fixed_acidity',
                           'citric_acid', 'residual_sugar', 'free_sulfur_dioxide', 'total_sulfur_dioxide']]

    targets = wine_set.quality_c

    pred_train, pred_test, tar_train, tar_test = train_test_split(predictors, targets, test_size=.4)

    # build model on training data#
    classifier = RandomForestClassifier(n_estimators=25)
    classifier = classifier.fit(pred_train, tar_train)

    predictions = classifier.predict(pred_test)
    # print the confusion matrix and accuracy of the model
    print('confusion matrix:\n', sklearn.metrics.confusion_matrix(tar_test, predictions))
    print('\naccuracy:', sklearn.metrics.accuracy_score(tar_test, predictions))

    # to display the relative importance of each predictive variable
    model = ExtraTreesClassifier()
    model.fit(pred_train, tar_train)

    print('importance of predictors:')
    dct = dict()
    for c in range(len(predictors.columns)):
        dct[predictors.columns[c]] = model.feature_importances_[c]
    print(sorted(dct.items(), key=operator.itemgetter(1), reverse=True))

    # run different numbers of trees to see the effect of the number on the accuracy of the prediction
    n = 100
    accuracy = [0]*n

    for i in range(n):
        classifier = RandomForestClassifier(n_estimators=i+1)
        classifier = classifier.fit(pred_train, tar_train)
        predictions = classifier.predict(pred_test)
        accuracy[i] = sklearn.metrics.accuracy_score(tar_test, predictions)

    plt.plot(range(1, n+1), accuracy)
    plt.xlabel("Number of trees")
    plt.ylabel("Accuracy of prediction")
    plt.title("Effect of the number of trees on the prediction accuracy")
    plt.show()

    print(accuracy)

# print('----------------Random Forests------------------------')
# call(random_forests)


# ________________________________Lasso Regression__________________________________