Research Of Collaborative Filtering Algorithm Based On User Preference And Trust Network

With the rapid development of information technology,a large amount of information on the Internet has brought people a wealth of information resources.In the age of information,people are satisfied with the large amount of data information which can meet their own needs,but what people confused about is the hidden useful information and the difficulty to get the real needs from those huge amounts of information.To solve these problems,personalized recommendation system arised at the historic moment.Personalized recommendation system can replace users to evaluate all the goods which they have not seen.It can help users find the items which they like or need according to their behavior and attributes.Recommendation algorithm based on collaborative filtering is the most widely used algorithm in the personalized recommendation system.It only needs user ratings,and has the characteristics of wide range of application,low limitation,high degree of personalization and can find the new points of interest easily.However,the collaborative filtering algorithm still faces the problem of accuracy,cold start and data sparseness.Based on the accuracy of the algorithm and the sparseness reduction of data,this paper first proposes a collaborative filtering algorithm based on user preference(UPCF).In view of the existing collaborative filtering algorithms do not take user's own rating criteria into account and ignoring the impact of time factors on user similarity and rating prediction,this algorithm considers the user's rating standards which would affect the user's similarity calculation and user preferences would change with time.And then,in order to alleviate the data sparsity problem,we propose a collaborative filtering algorithm based on user trust network(UTNCF)and solve communication problems of trust by adding a measure of trust value in the original algorithm and using the method of gain network flow.Also,we reduce the influence of data sparsity recommendation algorithm which based on user trusting degree to predict rating on the target user.Theory analysis and experimental results have shown comparing with the other t raditional collaborative filtering algorithms,that,UPCF algorithm can effectively im prove the accuracy of recommendation results in the same data set;In different sparse data sets,UTNCF algorithm can alleviate the impact of data sparsity and improve th e accuracy of recommendation results.