Improvements On An Algorithm For Ranking In Search Engine Based On Web Log Mining

The rapid development of Internet brings a great change to people; Information gives people not only happiness, but also some problems about how to use it. Now, the search engine is the most popular information retrieval tool. The personalized information service is much more attractive because the tradiontal search engine can't provide that service and there is too much resources on the Internet.In allusion to the need of search engine personlized service, the author puts forward the architecture of the personalized user interest model and the algorithms of each part of the system are analyzed and designed. at the same time, the author puts forward a new algorithm P_URank for ranking pages of a search engine based on personalized user model.The kernel thought of the improved algorithm in this thesis is exchanging a little efficiency for personalize service. In other words, we can greatly improve the service of our client by spending a little more time. Appraising Web pages not only depends on the hyperlink configuration between Webs, but also depends on consumer's valuation, and the final outcome not only depends on the pages'quality but also depends on a consumer's preference. In this way, the search engine can offer personalized service.At last, the author has confirmed the validity of the algorithmic by experiments, brings forward the direction of the next step of research and some potential problems.