Research On Collaborative Filtering Recommendation Algorithm Based On Trust Model

With the rapid development of Internet technology,user information and product information increased dramatically.Facing massive information,how quickly locate the information people need from these diverse information is a matter of concern.In order to solve the above information overload problem,researchers and engineers put forward a variety of technical solutions.The most widely technique is collaborative filtering recommendation system.This paper takes the trust factor in the social network as the object of study to solve the problem of data sparsity in collaborative filtering recommendation system and to improve the recommendation quality of recommendation system.We compare the traditional collaborative filtering recommendation algorithm and Trust-Aware Collaborative filtering in this paper.In view of the shortcomings of the traditional collaborative filtering algorithm and the existing Trust-Aware Collaborative filtering,the paper proposes two kinds of algorithms.The first one is the Implicit Trust Inference Method,named as ITIM,which is used to rebuild the trust rating matrix.The other is Implicit Trust-Aware Collaborative filtration,named as iTrace.The ITIM algorithm focuses on the transitivity and attenuation of trust in social networks,and calculates the missing value and the trust value of sparse trust rating matrices.The algorithm uses the Jeckard similarity coefficient to compute the similarity between the two user nodes in the Trust Network and adds the result as a trust value between the users from the trust matrix.Then,the trust value in trust network is iteratively updated by using Dijkstra algorithm,and finally the trust rating matrix is reconstructed.The ITrace algorithm combines the implicit trust inference algorithm model with the collaborative filtering algorithm model,to propose two scoring prediction methods,including linear weighted and increment weighted.In order to verify the validity and practicability of the algorithm,this paper uses the Filmtrust dataset to simulate the above algorithm and deploy the iTrace-I algorithm in a book recommendation system for testing.The results show that the research of implicit trust inference algorithm has a certain value in theory and application,and it also can greatly improve the recommendation quality of collaborative filtering recommendation system.