I created a module called sentiment_mod using NLTK. The classifier used for sentiment analysis is a custom classifier based on voting involving various algorithms including -
- Original Naive Bayes Classifier of NLTK
- Multinomial Naive Bayes
- Stochastic Gradient Descent (SGD)
- Bernoulli Naive Bayes
The module can do sentiment analysis on any piece of text and describe it as positive(pos) or negative(neg).
The file live_twitter_sentiment.py is used to do twitter analysis on live twitter feed on ay topic.
The file graphing_live_twitter.py plots a live graph of the sentiment value of tweets.
The pickle files can also be made by running the sentiment_module_1.py.