尝试使我的训练和测试数据的形状为n，1

Question

X_train, X_test, Y_train, Y_test =\
    train_test_split(shuffled_df.sentence_clean,shuffled_df.pol,test_size = 0.30,random_state=42)

tfidf1 = TfidfVectorizer(min_df=0.008, max_df=0.1)
#using tfidf to extract features as 
x_traincv = tfidf1.fit_transform(X_train)
x_testcv = tfidf1.fit_transform(X_test)

#Setting the polarities as integers
Y_train = Y_train.astype('int')
Y_test = Y_test.astype('int')

c = tree.DecisionTreeClassifier()

#Training the data
c.fit(x_traincv,Y_train)

＃问题在这里发生。我拼命试图计算模型的准确性。样本内和样本外错误。

training_accuracy = np.sum(c.predict(x_traincv) == Y_train)/float(Y_train.size)
testing_accuracy = np.sum(c.predict(x_testcv) == Y_test)/float(Y_test.size)

＃我得到的错误是：模型的特征数量必须与输入匹配。模型n_features是251，输入n_features是248

＃请让我知道是否有更好的方法。