Abstract: It is difficult to meet the individual requirements of users in the traditional ranking methods without considering the user preference,feedback and interest.In view of these problems,a new method of web page ranking is proposed based on web page similarity,link structure information,user preference and user feedback. Experimental results show that the improved ranking algorithm can improve the quality of web page retrieving and can meet the users’requirements greatly.

Key words: page ranking, search engine, link-analysis, user preference