I was wondering how one would calculate the pointwise mutual information for text classification. To be more exact, I want to classify tweets in categories. I have a dataset
