Research And Implementation Of Query Suggestion Model With Query Log And Corpus Data In Meta-search Engine

With the rapid development of information technology in 21 st century,the Internet has produced a huge variety of data,and the amount of data is still constantly increasing.At present,search engines can help people efficiently get the information they need from intricate and massive data.However,a single search engine often has low recall rate and could not fully meet the user's search needs.Thus,the meta-search engine integrates search results returned by various member search engines to provide search results with higher coverage for users.However,the original query words entered by the user may not accurately express their query intent in the meta-search engine,so that affects the accuracy of the search results.Therefore,it is very important to study the query recommendation technology of meta search engine,which could improve the user experience.The query recommendation technology recommend relevant query words to help the user construct more effective queries.The traditional query suggestion model can be divided into two categories,which are log-based query suggestion model and corpus-based query recommendation model.This paper first analyzes the advantages and disadvantages of these two types of models.Based on the “IM Search” meta-search engine,this paper studies and proposes a query suggestion model which can provide more accuracy candidate query words for users.The main work of this paper includes:(1)This paper first uses query log data to build a query-url bipartite graph,and uses a twostep random walk strategy to find the candidate query suggestion word set on the graph.In order to further expand the scope of the candidate word set,the search query suggestion results of multiple search engines are used to enrich the candidate query recommendation word set.Next,using the topic concept extraction method,the topic concepts related to the query words are extracted from the search results of the meta-search engine.And the candidate query suggestion words related to these topic concepts are filtered out.(2)In order to further improve the accuracy of the candidate query suggestion words,a similarity calculation method based on query item is proposed to obtain the final candidate query word set.In addition to this,this paper also proposes a method to calculate the rationality of query words,and sorts the candidate query word sets.(3)This paper validates the proposed query suggestion model.Experimental results show that the query suggestion model proposed in this paper can provide users with more accurate query suggestion words,and each step of the model can effectively improve the performance of query suggestion model.