朴素贝叶斯MultinomialNB ValueError：尺寸不匹配

Question

因此，我正在使用Naive Bayes MultinomialNB来预测Twitter情绪，我同时拥有训练和测试数据，并且我已经使用训练数据训练了该模型。然后我试图用测试数据进行预测，但它表示尺寸错误。我该如何解决？

getDisplayNameOfId(id:string):Observable<string>{
  let gdnid = new Subject<string>();

  firebase.database().ref("/users").orderByKey().once("value")
  .then((snapshot) => {
    let dni:string;
    snapshot.forEach((childSnapshot) => {
      if (childSnapshot.val().UserId === id){
        dni=childSnapshot.val().firstName + " " +childSnapshot.val().lastName;
      }
    });
    gdnid.next(dni);
  });


  return gdnid.asObservable();
}

因此，当我使用模型进行预测时，会引发尺寸错误。

from sklearn.naive_bayes import MultinomialNB
pf = pd.read_csv('/content/train.csv')
from sklearn.feature_extraction.text import TfidfVectorizer
tfidf_vectorizer = TfidfVectorizer()
tfidf = tfidf_vectorizer.fit_transform(pf.original_text)
model = MultinomialNB()
model.fit(tfidf,pf.sentiment_class)
test_csv = pd.read_csv("/content/test.csv")    
test_tfidf = tfidf_vectorizer.transform(test_csv['original_text'])

test_tfidf.toarray()

array([[0.        , 0.        , 0.        , ..., 0.15573603, 0.        ,
        0.        ],
       [0.2400925 , 0.        , 0.        , ..., 0.        , 0.        ,
        0.        ],
       [0.        , 0.        , 0.        , ..., 0.        , 0.46763784,
        0.        ],
       ...,
       [0.        , 0.        , 0.        , ..., 0.        , 0.        ,
        0.        ],
       [0.        , 0.        , 0.        , ..., 0.        , 0.182512  ,
        0.        ],
       [0.        , 0.        , 0.        , ..., 0.        , 0.42798355,
        0.        ]])