卡方检验提取特征来对文本分类

count_vect = CountVectorizer()  
X_train_counts = count_vect.fit_transform(corpus)     
tfidf_transformer = TfidfTransformer()  
X_train_tfidf = tfidf_transformer.fit_transform(X_train_counts)  
X_ch2 = SelectKBest(chi2, k=10)  
X_train_ch2 = X_ch2.fit_transform(X_train_counts, Y_train)  

#筛选之前的全部特征词
feature_list = list(count_vect.get_feature_names())
print(feature_list)

#输出查看一下筛选之后的特征词
idxs_selected = X_ch2.get_support(indices=True) 
list1 = []
for i in range(len(feature_list)):
    if i in idxs_selected:
        list1.append(feature_list[i])
print(list(set(list1)))

#预测
classifier = LinearSVC().fit(X_train_ch2, Y_train)