Research And Implementation Of Huatu Online Library System Based On Personalized Recommendation

Abstract:With the development of Information technology and Internet, the number of resources grows exponentially in network and we are entering the era of information overload from an era of scarcity gradually. Since more and more information are presented to users, it is hard to find the resources which they are interested in. In this situation, the internet is required to provide personalized recommendation service as soon as possible.With the investigation of the actual problem which the online library system faces in the personalized recommendation and the existed user collaborative algorithm, the thesis proposed a subclass collaborative recommendation algorithm based on distribution of user interest and also proposed an improved algorithm based on user time context. At last we introduced the user feedback mechanism in the actual system design.The thesis introduced the subclass of user interest, and then calculated the user’s score on subclass by analyzing the history data of user behavior operation. Based on the subclass of user interest, the algorithm considered the local similarity and calculated the user similarity^score prediction on a subclass rather than in overall. With the result of calculation, the algorithm could provide a final recommend list to user. Due to the user’s interest changes all the time in online library system the thesis introduced a time factor. Through analyzing the interests of different users at the same time, the algorithm could provide more accurate recommendation to target user. We also proved the result by doing experiment. Finally the thesis introduced the user feedback mechanism in the actual system design, further enhancing the accuracy of the algorithm and improving user experience.