Application Of Synonyms In Text Feature Extraction And Query Expansion

The rapid increase of modern information and the rising demand for high query quality, make it very difficult for traditional information retrieval technologies to meet users'needs, and some problems such as word matching and synonymous repetition in information retrieval will affect the improvement of retrieval efficiency to a certain extent.This paper studies the application of synonyms in text feature extraction and query expansion.In feature extraction, a method of improved TF-IDF feature extraction based on synonyms is proposed. The method adjustments the weights of synonyms and its related words based on TF-IDF through co-occurrence, and combines synonyms through the similarity.Adds the synonyms into local query expansion. At first, this paper presents a local query expansion based on synonyms method which uses original query terms to implement first query and get the top-ranked, gets synonymous of original query terms from this set, and then to expand user query. Then this paper improves local co-occurrence query expansion, which utilizes the related words of original query terms and its synonyms in the top-ranked documents as expansion terms.The above methods makes experimental verification, experimental results show that the importance of synonyms can't be ignored in text feature extraction and query expansion.