训练模型失败，因为'list'对象没有属性'lower'

Question

我正在通过推文训练分类器以进行情感分析。

代码如下：

df = pd.read_csv('Trainded Dataset Sentiment.csv', error_bad_lines=False)
df.head(5)

#TWEET
X = df[['SentimentText']].loc[2:50000]
#SENTIMENT LABEL
y = df[['Sentiment']].loc[2:50000]

#Apply Normalizer function over the tweets
X['Normalized Text'] = X.SentimentText.apply(text_normalization_sentiment)
X = X['Normalized Text']

规范化后，数据框如下所示：

X_train, X_test, y_train, y_test =
sklearn.cross_validation.train_test_split(X, y, 
test_size=0.2, random_state=42)

#Classifier
vec = TfidfVectorizer(min_df=5, max_df=0.95, sublinear_tf=True,
                      use_idf=True, ngram_range=(1,2))
svm_clf = svm.LinearSVC(C=0.1)
vec_clf = Pipeline([('vectorizer', vec), ('pac', svm_clf)])
vec_clf.fit(X_train, y_train) #Problem
joblib.dump(vec_clf, 'svmClassifier.pk1', compress=3)

失败并出现以下错误：

AttributeError: 'list' object has no attribute 'lower'

Full Traceback:
--------------------------------------------------------------------------- AttributeError                            Traceback (most recent call last) <ipython-input-33-4264de810c2b> in <module>()
      4 svm_clf = svm.LinearSVC(C=0.1)
      5 vec_clf = Pipeline([('vectorizer', vec), ('pac', svm_clf)])
----> 6 vec_clf.fit(X_train, y_train)
      7 joblib.dump(vec_clf, 'svmClassifier.pk1', compress=3)

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\pipeline.py in fit(self, X, y, **fit_params)
    255             This estimator
    256         """
--> 257         Xt, fit_params = self._fit(X, y, **fit_params)
    258         if self._final_estimator is not None:
    259             self._final_estimator.fit(Xt, y, **fit_params)

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\pipeline.py in
_fit(self, X, y, **fit_params)
    220                 Xt, fitted_transformer = fit_transform_one_cached(
    221                     cloned_transformer, None, Xt, y,
--> 222                     **fit_params_steps[name])
    223                 # Replace the transformer of the step with the fitted
    224                 # transformer. This is necessary when loading the transformer

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\externals\joblib\memory.py in __call__(self, *args, **kwargs)
    360 
    361     def __call__(self, *args, **kwargs):
--> 362         return self.func(*args, **kwargs)
    363 
    364     def call_and_shelve(self, *args, **kwargs):

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\pipeline.py in
_fit_transform_one(transformer, weight, X, y, **fit_params)
    587                        **fit_params):
    588     if hasattr(transformer, 'fit_transform'):
--> 589         res = transformer.fit_transform(X, y, **fit_params)
    590     else:
    591         res = transformer.fit(X, y, **fit_params).transform(X)

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\feature_extraction\text.py in fit_transform(self, raw_documents, y)    1379             Tf-idf-weighted document-term matrix.    1380         """
-> 1381         X = super(TfidfVectorizer, self).fit_transform(raw_documents)    1382         self._tfidf.fit(X)  1383         # X is already a transformed view of raw_documents so

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\feature_extraction\text.py in fit_transform(self, raw_documents, y)
    867 
    868         vocabulary, X = self._count_vocab(raw_documents,
--> 869                                           self.fixed_vocabulary_)
    870 
    871         if self.binary:

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\feature_extraction\text.py in _count_vocab(self, raw_documents, fixed_vocab)
    790         for doc in raw_documents:
    791             feature_counter = {}
--> 792             for feature in analyze(doc):
    793                 try:
    794                     feature_idx = vocabulary[feature]

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\feature_extraction\text.py in <lambda>(doc)
    264 
    265             return lambda doc: self._word_ngrams(
--> 266                 tokenize(preprocess(self.decode(doc))), stop_words)
    267 
    268         else:

C:\Users\Monviso\Anaconda3\lib\site-packages\sklearn\feature_extraction\text.py in <lambda>(x)
    230 
    231         if self.lowercase:
--> 232             return lambda x: strip_accents(x.lower())
    233         else:
    234             return strip_accents

AttributeError: 'list' object has no attribute 'lower'

Answer 1

TFIDF Vectorizer应该期待一个字符串数组。因此，如果你向他传递一系列tokenz数组，它就会崩溃。

Answer 2

答案来自 http://www.davidsbatista.net/blog/2018/02/28/TfidfVectorizer/

from sklearn.feature_extraction.text import CountVectorizer

def dummy(doc):
    return doc

tfidf = CountVectorizer(
    tokenizer=dummy,
    preprocessor=dummy,
)  

docs = [
    ['hello', 'world', '.'],
    ['hello', 'world'],
    ['again', 'hello', 'world']
]

tfidf.fit(docs)
tfidf.get_feature_names()
# ['.', 'again', 'hello', 'world']

Answer 3

将此代码.apply(lambda x: ' '.join(x))添加到X_train和y_train之后，它应该可以工作。