Earticle

Recently, with the rapid development of science and cultural areas, the written and electronic amount of books in the university libraries increases sharply. When readers use the traditional searching system which is based on the collaborative filtering method, it is usually difficult for them to find out their interested books due to a large number of results from the system. Aiming at this problem, the paper points out a sort of personalized recommending system. This system optimizes the collaborative filtering method based on the information of the users, uses the new collaborative filtering method based on the classification of users and books, and analyzes and orders a variety of recommending information after filter. From the experiment, it concludes that compared with the traditional recommending system the personalized system targets towards different types of readers. The numbers of books which feed back to readers have decreased a lot. What’s more, after a comprehensive analysis and order of the various recommending information, the average recommending accuracy will make further improvement.