Abstract:

The text retrieval with keywords is analyzed, and the recall rate is low without the expanding query. And as for the text retrieval with concept, the weights between the expanding query terms and query words are not distinguished. So, a method named  expansion lessening, which is used for computing the weight of the expanding query term by the dint of words semantic similarity, is proposed. By expansion lessening, the expanding query term and query words are used to construct a  vector space model for  text search. Experiments show that the integrated performance of the constructed retrieval model is improved greatly in text search.

Key words: expanding term;weight calculation;text retrieval