摘要: Query expansion methods have been studied for a long time - with debatable success in many instances. In this paper we present probabilistic query model based on similarity thesaurus which was constructed automatically. A reflects domain knowledge about the particular collection from it is constructed. We address two important issues expansion: selection and weighting of additional search terms. contrast to earlier methods, our queries are expanded by adding those terms that most similar concept query, rather than selecting Our experiments show kind results notable improvement retrieval effectiveness when measured using both recall-precision usefulness.