Improvement on corpus-based word similarity using vector space models
Yunus Emre Esin, Özgür Alan, Ferda Nur Alpaslan · 2009
This paper presents a new approach for finding semantically similar words from large text collection using window based context methods. Previous studies on this problem mainly concentrate on finding new methods which are new combination of distance-weight measurement methods or new context methods. The main difference of our approach is that we focus on reprocessing of existing methods' outputs to update the representation of related word vectors, which are used for measuring semantic distance between words, to further improve the results. This new approach can be easily applied to many of the existing word similarity methods using the vector space model for representing contexts. We claim that our method improves the performance of some of the existing similarity measuring methods.