TextBlob Naive Bayes classifier for neutral tweets

Question

I am doing a small project on sentiment analysis using TextBlob. I understand there are are 2 ways to check the sentiment of tweet:

1. Tweet polarity: Using it I can tell whether the tweet is positive, negative or neutral
2. Training a classifier: I am using this method where I am training a TextBlob Naive Bayes classifier on positive and negative tweets and using the classifier to classify tweet either as 'positive' or 'negative'.

My question is, using the Naive bayes classifier, can I also classify the tweet as 'neutral' ? In other words, can the 'sentiment polarity' defined in option 1 can somehow be used in option 2 ?

Answer 1

If you have only two classes, Positive and Negative, and you want to predict if a tweet is Neutral, you can do so by predicting class probabilities.

For example, a tweet predicted as 80% Positive remains Postive. However, a tweet predicting as 50% Postive could be Neutral instead.