The Research Of Query Expansion Based On Buffer Strategy

With the rapid development of Internet, the amount of online information is growing quickly. The technology of information retrieval solves the problem that gets the needed information from the mass of Internet data. However, the information retrieval technology has some limitations. It has not solved the problems such as words mismatched between the user's querying sentence and the documents. In order to solve the disadvantage of information retrieval, the academic community proposed the method of query expansion, which can well improve the performance of the retrieval system.The technique of local co-occurrence computes the co-occurrence rate between the index terms and the query of the feedback documents. The buffer strategy for query expansion is based on the method of local co-occurrence. This method saves the co-occurrence information of the query terms and the extended words, which is obtained in the extended word matrix. For the user queries, the expansion of the query word is expanded by the extended word matrix. Thus the number of expansion words is expanded.The processing of large number of information data in query expansion system will become significantly slower. The buffer strategy for query expansion caches the required documents. When the system retrieves the documents, if the buffer system contains copy of the documents, the copy can be directly feedback to the user and improve the speed of retrieval system.The query expansion method based on buffer extended words and the query expansion based on the buffer strategy is combined. The result is improved and the system performance is improved.