Sentiment analysis with NLTK and Scikit-learn sklearn

sentiment_test.py

# https://twitter.com/realDonaldTrump/status/1309268149242597377
tweet = 'We are providing better care, and more choice, at lower cost. We are delivering a healthier, safer, brighter, and more prosperous future for EVERY citizen in our magnificent land – because we are proudly putting AMERICA FIRST!'
preprocess_tweet = preprocessing(tweet)
print(preprocess_tweet)

['provide', 'better', 'care', 'choice', 'lower', 'cost', 'deliver', 'healthier', 'safer', 'brighter', 'prosperous', 'future', 'EVERY', 'citizen', 'magnificent', 'land', 'proudly', 'put', 'AMERICA', 'FIRST']

# vectorize the tweet
tweet_vec = vec.transform(pd.Series([tweet]))

# predict a label
tweet_prediction = clf_rfc.predict(tweet_vec.toarray())
tweet_prediction = 'positive' if tweet_prediction[0] == '1' else 'negative'
print('{} has been predicted for the tweet {}'.format(tweet_prediction, tweet))

negative has been predicted for the tweet We are providing better care, and more choice, at lower cost. We are delivering a healthier, safer, brighter, and more prosperous future for EVERY citizen in our magnificent land – because we are proudly putting AMERICA FIRST!