Stemming normalizes words: Porter stemming It's language dependent: snowball Several algorithms exist pecl/stem arrival -> arrive skies -> sky riding -> ride rides -> ride horses -> hors Alternatively, instead of word analysis you can use "sounds like" indexing, by using something like soundex or metaphone: Word Soundex Metaphone stemming S355 STMNK peas P200 PS peace P200 PS please P420 PLS