This is a work in progress
I am working on a perfect simple parser that can replace the bag-of-word approach
so far I have a simple parser that can take sentences and break them into word frequency lists don't believe will become --> don't:1 believe:-1