# Feature engineering of closely related text

I am trying to do multi class classification of text. For many reasons I can't paste the data, atleast now in open.

The problem is there is a text of closely related subjects like Anatomy and physiology. There are other categories like dental, Ortho for which I am able to classify correctly. I have used Countvectorizer, Tf-IDF with MultionialNB of scikit-learn.

I get good results for dental and ortho because they donot have many common words. However, the error rate is high between Anatomy and Physiology categories as they have lot of common words.

What's the best way to proceed with feature engineering for this problem?