Research On Collaborative Filtering Algorithm Based On Trust Relation And Interest Change

With the rapid development of Web2.0 technology and information communication technology,we have changed from the era of information scarcity to the era of information overload.Personalized recommendation technology is the effective method to solve the problem of information overload has attracted extensive attention of researchers,among them,the collaborative filtering algorithm is considered to be one of the most popular research recommendation algorithm,but it still has a lot of space for improvement.For example,the traditional collaborative filtering algorithm can not effectively resist the attacks of malicious users,resulting in a decrease in the accuracy of recommendation.As well as the problem of information expiration,even if there are a lot of researchers put forward time decay function can effectively alleviate the problem,but the accuracy of the recommendation is not ideal enough.In this paper,we study the existing trust model and the time decay function from the point of view of user trust and interest change,the main work is as follows:1.Existing collaborative filtering algorithms based on trust relations usually measure user trust with user interaction,However,the simple calculation of interaction behavior in the current trust measure weakens the influence of the difference of user’s individual preference on interactive behavior.Therefore,this paper introduces the user preferences based on traditional trust model,and presents an improved collaborative filtering algorithm based on trust relations.The experimental results show that this method can improve the accuracy of recommendation.2.To some extent,the collaborative filtering algorithm which is integrated into the traditional time decay function can depict the change of information value over time,But it is not considered that the change of user’s interest is invariant in local time.therefore,the traditional time decay function can be further optimized.Under the guidance of the time window,the user interest invariant time window is introduced into the time decay function..Finally,an improved collaborative filtering algorithm based on interest change is proposed to improve the accuracy of recommendation.3.On the basis of the above work,considering the trust relationship with the user interest change two factors,using the weighted mixing these two factors into collaborative filtering algorithm,this paper proposes a collaborative filtering algorithm based on hybridtrust relationship with the interest change.The experimental results show that the recommendation algorithm combined with the two factors has better recommendation effect than the single model.In summary,this paper aims at the problem that the collaborative filtering algorithm can not effectively resist malicious user attacks and outdated information.On the basis of existing trust model and the time decay function,user preferences and interest invariant time windows are introduced respectively,this paper proposes a collaborative filtering algorithm of mixed trust relationship with the interest change.The results of this study will help to provide users with more reliable and more accurate recommendation services.