Application Research Of Collaborative Filtering Technology For The Library Personalized Service

With the rapid development of Internet technology and the rise of e-commerce market,it provides a good environment for the development of personalized recommendation technology.With the great success of personalized recommendation technology in e-commerce market,there is relatively less on the application of personalized recommendation technology in the field of library service.With the overall promotion of national education,people want to achieve more knowledge to enrich themselves,and library provides a platform for people to learn new knowledge.However,with the rapid development of knowledge,it is difficult for people to find books which accord with their own preferences from the vast resources of books.So it has become an great research significance to use personalized recommendation technology in library service by learning from the successful experience in the field of e-commerce market and the collaborative filtering recommendation algorithm is the most widely applied in many recommendation methods.How to apply collaborative filtering recommendation technology in the personalized service of Library is this dissertation's research emphasis,organized as follows:(1)This dissertation first analyzes the research status of personalized recommendation system for library,and then focuses on a variety of personalized recommendation systems and compares their advantages and disadvantages.(2)In this dissertation,an improved collaborative filtering recommendation algorithm is proposed based on the characteristics of the library recommender system.Firstly,the background information and users' ratings of the items are combined to calculate the similarity between users,this method can calculate the similarity of users more accurately and avoid the problem of cold-start;Secondly,it uses hierarchical clustering method to cluster users and similar user preference groups are constructed,the change trend of the interest preference for the user group with time are described by project-label network diagram and label-time network diagram of the similar user preference groups.Finally,it generates recommendations for target users.We conducted experiment to verify the efficiency of the algorithm and the comparison to other recommendation algorithms by publicly available data sets;the experimental results indicate that this method can effectively solve the cold-start of the personalized recommendation system,the problem of low efficiency.(3)The improved collaborative filtering recommendation algorithm is used to design a prototype of the library recommender system in this dissertation.