I want to make a program that can classify news into real or fake news. In this project I've used Naive Bayes Algorithm to solve it. I have categorized fake news with sentimental analysis and I already know how to classify it.
But now I want to add constraint in this project, which is I want to have punctuation amount to decide if its a real or fake news but I am confused how to mix it. In the like-hood or what?
Thank you. Before it I have already used n-gram and laplace smoothing in the sentimental analysis of the news.