You'll be wanting a minimum of a naive stemming algorithm (test the Porter stemmer; you can find readily available, free code in the majority of languages) to system textual content initially. Retain this processed text plus the preprocessed textual content in two separate Room-break up arrays.