Research On Social Recommendation Method Considering Group Information

With the rapid development of Web2.0 technology,a large number of user generated content oriented social media applications have emerged on the internet.making users not only the recipient of internet content,but also become a manufacturer.However,with the rapid growth of users participating in social media,social media gathered a large number of users generated data,which results in serious information overload problem,making it is difficult for users to find information related to their requirement.The recommendation system is one of the effective ways to solve the information overload problem and realize personalized service.Therefore,the recommendation system research for social media,social recommendation system gradually received more and more attention from academic and industry.Currently,many researchers have carried out relevant research on social recommendation system,and proposed some social recommendation methods.However,these existing social recommendation methods have not fully exploited and utilized the group information on social media.While group information as a more open social relationship on the social media,its unique social attributes have an important influence on social recommendation method,such as group size,group composition,group structure and so on.Therefore,this research proposed a new social recommendation method considering group information to solve the above problems.Firstly,research status on social recommendation system was analyzed,which clarifies the research problems and future directions.Secondly,the basic theory of social recommendation system,including the concept of social recommendation system,the framework of social recommendation system,the common social recommendation methods for individual and group was analyzed.Then,based on the above analysis,this research from the perspective of individual and group,proposed a social recommendation method considering group information for individual,and a social recommendation method based on unified probabilistic matrix factorization for group.Finally,in order to verify the validity of the proposed method,this research crawled data from a chinese social media called Dou Ban and a famous scientific social network called CiteULike to conduct experiments.Experimental results show that the proposed method gets the best recommendation results,and proved the effectiveness of the new proposed method.Through this research,on the one hand,the relevant theory of social recommendation system was sorted out systematically,based on this,an improved social recommendation method was proposed,which enriched and expanded the theory of social recommendation system.On the other hand,this research analyzed the group information on social media,and integrated it into social recommendation method,which has expanded the scope of social information utilization in social media,and provided a new way for social recommendation methods integrate other information.