Query Expansion Based On The Association Rules

With the wealth of Internet information resources continually, when users retrieve information using a search engine when a lot of useless information will be feedback link, so search engines can not meet the existing one for the speed and accuracy requirements. As users browse the web logs record the behavior of users to access web pages and web information, a direct reflection of the rules and user interest in using the network, through mining user browsing the web log and its contents, you can extract the user's interest, according to the user's interest , the establishment of the configuration information file describes the user in the Web search engine users search, the reference to the user's interest model, the user submits a query term correction and query expansion to improve the accuracy of query words described, and the query results based on users use the network and the user interest of association rules to re-sort information in order to improve the efficiency of user queries.In this paper, on the Web log mining, association rule mining, query expansion, and other related principles and technical analysis, based on in-depth analysis of the client user network log mining and associated user interest model based extraction of key technologies such as mining, concrete results are as follows :Through the development of IE plug-in, the client collecting user behavior and browsing the web log information to analyze the user's Web log, used to mine user access to the network of users access the Web page URL excavation analysis, extracts the user's interest in class and at the same time URL to Web content according to the description of the characteristics of its classification of web content for the same class cluster analysis, classification of sets of user interest mining and feature extraction.Features based on user interest and user interest set, build the tree based on user interest model, based on user interest model, the user is browsing the association rules mining, mining the frequent itemsets.Model based on user interest and user interest in the phrase feature set, users use a search engine for information retrieval when the query words for query expansion, search engine results based on user feedback, visit the website of the association rules to re-sort the frequent items to improve the efficiency of the user query information.