python - X 每个样本有 19257 个特征；期待 19234，同时使用 Logistic 回归 Pickle 模型进行预测

Question

当我试图预测来自泡菜的模型时，我正面临这个问题

from sklearn.feature_extraction.text import TfidfVectorizer

from sklearn.feature_extraction.text import CountVectorizer

recreated_vec = CountVectorizer(decode_error ='replace' ,vocabulary = vocab)

from sklearn.feature_extraction.text  import TfidfTransformer

transformer = TfidfTransformer()

recreated_vec.fit_transform(df) 

(output: <54045x19257 sparse matrix of type '<class 'numpy.int64'>'
    with 1086100 stored elements in Compressed Sparse Row format>)


model.predict(transformer.fit_transform(recreated_vec.transform(df)))

ValueError: X has 19257 features per sample; expecting 19234

谁能帮我解决这个问题？

标签： pythonpicklelogistic-regressionpredict