I’m happy to share my project on “Build a SMS spam filter with Naive Bayes”.
The first part of the project pretty much follows instructions from the guides. After that, I revised the project (you can skip to Revision 1 section) to take special characters and upper case into account. At the end, I also try to use built-in sklearn Naive Bayes package such as: MultinomialNB, GaussianNB, ComplementNB to check if the results got from sklearn match with out implementation-from-scratch.
I appreciate any feedback!.