Research On Some Key Technologies Of Personlized Meta Search

The average precision of some famous Search Engines is not high enough currently. What's more, different Search Engine gives you different result. So if one wants to get more comprehensive, veracious retrieved information, he must use lots of Search Engines. Then the Meta-Search engine system has been designed and resolves this problem in some way and this give Meta-Search Engine a rapid development. In addition, everyone's interests are not the same, as what they emphasized on. Existing search engine can't satisfy users for they can only return the same results for everyone's search. So, personalized search becomes the hot point of information retrieval research.The meta search engine adopts many search engines'query ability to enlarge the scope of searching. The personalized search can satisfy the user's personalized need through providing the most relevant information. A personalized meta search engine combines the advantages of these two technologies and provides an effective method to improve the search engine's querying ability.The state-of-the-art of meta search engines is firstly overviewed as well as that of personalized search. A summary of meta search and personalized search's status quo is made. The review of interrelated theories and algorithms are researched. Then the analyses and comparisons of personalized search theories and techniques that have been proposed for personalized meta search engines are provide. Afterwards, based on the researches, we propose several algorithms for the implement of personalized meta search. Focus on the user's personalized demand, a user interest model based on classified hiberarchy model is established. And, three algorithms on establish and update of user interest model are proposed. Focus on the result merging, an algorithm of results merging based on Borda Function of decision-making theory is proposed which can merging the results from member search engines well and satisfied the user demands.Finally,design and implement of a prototype system are explained. And on it, a variety of experiments have been carried out. The analysis of experimental data indicates our methods are effective.