Combining Query Logs And Ontology For Query Optimization

The user input query statement is short,which leads to the information retrieval system can not calculate accurate answers.In order to solve this problem,this article mainly sets forth that we can execute enlarging query language in the method of converged network log and ontology,while ontology is completed on the basis of message retrieval from wiki through property,gradation,synonym and semantic similarity.In a Wikipedia-based query expansion study,we firstly extract 8,000,000 subject terms from Wiki and then acquire expansion words three aspects,respectively,property,gradation,synonym.There is amounts of dependencies in Wiki,which provide the possibility that make full use of relationships between them.In conclusion,we will obtain expansion words from property,gradation,synonym and semantic similarity.In the converged network log query extension,I illuminate that how to get the optimal extension word by three steps:clustering the query words by K-Means++,dealing with the corresponding web page document by LDA topic model and calculating.Next we can manipulate the expansion words through user log and finally obtain the definite expansion words to be added in the primary query.In all,this banishes the shortness of query drift and improve the efficiency of query system.This paper have two Innovation:1,Improve the systemic recall on the basis of property,gradation,synonym and semantic similarity.2,Analyze user log in detail through LDA topic model,K-Means++ and specific data segment.To solve the problem of query drift and improve the efficiency of query system,we execute the secondary management...