Personalized Recommendation Methods Based On Weak Ties In Social Networks

In recent years,with the development of mobile Internet,people begin to change from the receiver of network information to the producer of information.Social network is a typical product with the mobile Internet,people can communicate more conveniently,people are very fond of this way of new social.On social networking sites,people can communicate with their friends.In the era of mass data,the amount of information is increasing,so the friends recommendati on is important in social networks.More and more researchers have begun to study the friends recommendation algorithm.Existing recommendation algorithms are often based on users ' attributes or interests.Few studies have considered the users' information benefits,and the network information circulation.Therefore,this paper proposes a friends recommendation algorithm based on weak ties.The research in this paper mainly focuses on weak relationship.This paper first defines the strong and weak relationships in the network,according to the sociological theory.This paper selects the community discovery algorithm as the method of dividing the strong-weak relationship.The original evaluation criteria of the recommendation algorithm are no longer applicable to the evaluation of weak relationship recommendation.So this paper proposes a new evaluation standard to evaluate the ability of acquiring heterogeneous information of nodes.At the same time,this paper measures the effect of the classical recommendation algorithm on the new standard.According to weak relation theory and structure hole theory,a new weak tie recommendation algorithm is proposed in this paper.Experiments show that the weak tie recommendation algorithm has better heterogeneous information acquisition ability than the classical recommendation algorithm,which indicates that the weak tie recommendation algorithm is effective.Besides,this paper also improve the classical recommendation algorithm by using the weak tie recommendation algorithm.