Analysis And Research, Personalized Information Retrieval Based On User Interest

With the development of the computer technology and the needing of Internet, Search Engine appears for people to get access to the global information service exactly and in good time. As the most basal means for network information service, Search Engine may satisfy the user to a certain extent to WWW on the information retrieval request. While information content enormous rich, the user faces the question of "the overloading information" and "the confusing resources". Solely relying on traditional search engines has not been able to meet the demand of the People, the personalized search engine"based on user interest"becomes the main method of solving the above question, the personalized search engine and the personalized service technology has become the hot spot of the present research .In this paper, we combine user interest modeling technology with query expansion methods in the personalized search engine, which can improve the personalized serviceability of the personalized search engine. The content of this paper is as follows:Firstly, this paper proposes that dendriform vector space model express user interests. Because it is not accurate or comprehensive for vector space model to express user interests and the VSM doesn't consider multiformity of user interests. The experiment proves that it can be accurate and comprehensive for dendriform VSM to express user interests. Through analyzing the behavior of the user in using search engine, this paper puts forward a method of computing the page interest degree based on page browsing speed. The experiment proves that the result computed based on this method is close to fact value.Secondly, in order to enable the user interest information to adjust automatically and the auto-adapted dynamic change along with the user interest change, we need to adjust the user interest characteristic word auto-adapted the weight and the refresh time information. In the user interest model's renewal, we uses one week as a circulation. It causes the user interest model to descript user's interest hobby more accurately.Thirdly, we make research on the methods of query expansion basing on the user feedback and analyze the defects of the traditional and present methods of query expansion basing on the user feedback. After that we amend the method of query expansion basing on the user feedback and research the related adjustment parameters.Finally,this paper discusses the construction of personalized Search Engine, and gives the designs of every function modules, which combine the user interest model and the query expansion to satisfy the personalized demand. Through the experimental verification, personalized search has a higher accuracy ratio than Ordinary inquiry.