Font Size: a A A

Result Optimization Of Meta Search Engine Based On Agent

Posted on:2008-11-23Degree:MasterType:Thesis
Country:ChinaCandidate:H WangFull Text:PDF
GTID:2178360218463606Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
At present, among all the search engines, no one can cover all the WWW resources. Most of them can only cover a very small part of the entire resource. The difference among information sources of all kinds of search engines is so large that Meta Search Engine (MSE) appears which integrates many independent search engines. MSE has larger cover face, better search efficiency and expansibility. Result optimization is a key technology in aspect of improving meta-search engine's efficiency.It is a remarkable problem that the amount of results returned by search engines is enormous and engines can not provide to users with required results based on their interests. By researching on the clustering technology of web page, with concept extraction as goal, a real-time clustering algorithm was brought forward. Different from the traditional clustering strategy, it can not only classify the result document sets but also give its topic description effectively with the advantages of lower complexity and higher flexibility. Through analyzing result integration algorithms, we proposed a kind of mapping algorithm from user's browsing behavior to user's interests and improved topic-sensitive page rank algorithm combining with user's interests. It realizes the personal compositor of search results and figures out the problem that the amount of results returned by search engines is too large and usually users don't have enough patience to look through all the results. As a result, this algorithm improves the query quality and efficiency obviously. At last, based on the research of Agent and Multi-Agent technology, this paper designed a testing system for meta-search engine based on Multi-Agent and proposed a multi-agent system framework based on JADE which accomplished the parallel, distributional processing of information collection process. It is the system not the users that draws comparisons among the search results, removes redundancy web pages and returns to users with web pages in good order. Contrasted to traditional object-oriented system, it has quicker response speed and lower coupling.
Keywords/Search Tags:Meta Search Engine, Agent, result optimization, subject classification, user's interes
PDF Full Text Request
Related items