| With the developing of Internet, the traditional serve pattern of the people find information" has become more and more difficult to adapt the information developing fast. Web personalization emerges as the times require. The Web personalization is the process of customizing a Web site to the needs of the specific users, the taking advantage the of the knowledge acquired the from the analysis of the user' the s navigational the behavior (the usage data) in the correlation with the other information the collected in the Web context, namely, structure, content, and user profile data. Existing Web personal service contains three kinds of forms: personal recommendation, personal Web site, and personal information search. There ware many personal service systems already applied in the in the electronic commerce, E-mail, digital library et al. But the personal service technique still has many worthy of research and improvement of place.This paper carries on the thorough research to the personal recommendation in analyzing the localization of the existing technique and the state of domestic and international research. Paper discusses the three key technique come down to the personal recommendation. That is user modeling (user profiling), Web resource profiling, and recommendation. At the part of user modeling, Web log mining and a method of data preprocess are introduced. For the Chinese word segmentation which involved in Web resource profiling, paper introduced a segmentation method of bidirectional matching, and use the segmentation of no dictionary based on statistical to the process of the phrase which did not register. Paper also propose a new method of describe Web pages which use the attribute classification based on the research of text character abstract and expert system. This made the description of Web pages more complete and benefit in calculation. For recommendation, paper discussed the method of mix recommendation which combine the content-based and the collaborative filtering. It not only meet the interest that user have been but also discover the new interest of user. Above the research, paper proposes a structure model of personal information service. On the analysis and mining of user information, it recommend to user according to they hobby. In addition, the collection of user information is inexplicitly which dose not demand user provide any information. On the one hand, conquer the problem of user did not provide the individual information so can not obtain the avail information. On the other hand, it protects the user privacy. Finally, paper release a personal recommendation system of movie Web combining the typical application field, and put forward the research direction of the aftertime. |
