Research On Trust Relationship Recommendation Algorithm Combining Potential Factor And Time Factor

The rapid development of Internet and information technology has greatly promoted the exponential growth of user generated content(UGC).As an independent individual,users participate in social activities,analyze the interactive data of user trust,and make personalized trust recommendation for users.In view of the problems existing in the research on the algorithm of trust recommendation,such as data sparsity,cold start,and insufficient information mining,this paper clearly defines the concepts of trust experience and trust level and proposes the corresponding solutions by integrating two factors affecting trust such as potential factor,time factor.This paper first analyzes the research status of trust recommendation algorithms.For the problem of low recommendation accuracy caused by sparse data and cold start,this article will use the latent factor model(LFM)to analyze the impact of user-item potential factors through dimensionality reduction.Leverage iterations to adjust the user and item potential vectors,and combining trust influencing factors of the users,a specific explanation of the potential vectors of each dimension is given.Secondly,for the problem of insufficient recommendation of trusted user-item information,analyze the influence of trust experience and trust degree of users on their behavioral interest in both explicit and implicit trust.At the same time,the dissemination and asymmetry of trust are considered,the user’s potential trust and interest preference patterns are explored,and a trust relationship recommendation algorithm based on potential factors is proposed.Thirdly,for the problem of trust value changing in the recommendation process is addressed,without giving a specific explicit trust,through user-item data mining implicit trust,time factors are taken into account while user behavior and interest are analyzed,and introduce time factor,explore the trust change trend of users,and apply it to trust recommendation,further explore the time pattern of user behavior and interest,and a trust relationship recommendation algorithm that fuses potential and time factors is proposed.Finally,the two algorithms proposed in this paper are experimentally verified on real data sets.