新闻报告的文本分类。数据集将1000个文档分成20个类。要求：五折交叉验证结果
    数据集：http://www.cs.cmu.edu/afs/cs/project/theo-11/www/naive-bayes.html

    1. from sklearn.naive_bayes import MultinomialNB
    2. from sklearn.feature_extraction.text import TfidfVectorizer
    3. from sklearn.datasets import fetch_20newsgroups
    4. from sklearn.model_selection import StratifiedKFold
    5. import numpy as np
    7. def fold_5(news_data):
    8.     x = np.array(news_data.data)
    9.     y = news_data.target
    10.     skf=StratifiedKFold(n_splits = 5)
    11.     skf.get_n_splits(x,y)
    12.     for train_index,test_index in skf.split(x,y):
    13.         x_train,x_test = x[train_index],x[test_index]
    14.         y_train,y_test = y[train_index],y[test_index]
    15.         naive_bayes(x_train,x_test,y_train,y_test)
    17. def naive_bayes(x_train,x_test,y_train,y_test):
    18.     #进行特征抽取
    19.     tf = TfidfVectorizer()
    20.     x_train = tf.fit_transform(x_train)
    21.     x_test = tf.transform(x_test)
    22.     #进行朴素贝叶斯算法分类
    23.     bayes = MultinomialNB(alpha=1.0)
    24.     bayes.fit(x_train,y_train)
    25.     y_predict = bayes.predict(x_test)
    26.     print("测试集的预测结果为：",y_predict)
    27.     print("模型的预测准确率为：",bayes.score(x_test,y_test))
    29. if __name__ == '__main__':
    30.     news_data = fetch_20newsgroups(subset="all")    #读取数据
    31.     fold_5(news_data)   #五重交叉验证