A Research On Personalized Recommendation Methods Based On Trust Propagation And Singular Value Decomposition

With the era of big data coming,the Internet is filled with a large number of complex data,how to quickly and efficiently find the right information become one of the hot spots of concern.Recommender system is a kind of technology which is based on the information filtering technology to recommend the interested information for user.The emergence of the recommended system effectively alleviates the problem of information overload.However,the traditional recommendation system still has the following challenges:First,most of the traditional recommendation algorithm assumes that users are independent of each other,this assumption ignores the impact of social relationships between users on user decisions,which is not consistent with our social relationships in the real world.Second,although some recommendation systems have begun to focus on the trust relationship between users,trust information is also very sparse,this will lead to most of datasets only contains very little information about the user's relationship.In this paper,we proposed an innovative method that integrated users' trust propagation and singular value decomposition into recommendation Algorithm to improve the quality of the recommendation effectively and efficiently.First of all,aiming at the traditional recommendation algorithm ignores the social relations among unknown users,we proposes a trust propagation algorithm.This algorithm figure out the indirect trust between users based on the direct trust,and then to fill the user trust matrix,then alleviated the sparse problem of the trust relationship matrix,improved the recommendation system recommended quality.Secondly,due to the sparseness of the user's score matrix,the recommendation quality of the recommendation system is reduced.In this paper,we user the singular value decomposition model.The SVD model can map the data to the low-dimensional space,and then calculate the similarity between the items in the low-dimensional space,evaluate the user's unrated item,and finally recommend item with high forecast score to user,improves the recommended quality of the recommended system.In this paper,it is helpful to improve the recommendation quality of the recommendation system by combining the trust propagation rules and the singular value decomposition model.Finally,we performed our experiments on two real data sets respectively,the public domain Epinions.com and Filemtrust.com.The experimental results show that our method has a better outperform.