Research On Social Circles Discovery Algorithm Of Personal Network

With the continuous development of the Internet,online social platform rapidly grow,such as domestic QQ,WeChat,foreign Facebook,Twitter and so on.The application and popularization of these social platforms has changed the way of people’s communication,expanded their social network,and made the prospect of social networking analysis and mining more and more broad.The social circles discovery of personal network is an important aspect of social network analysis and mining.It is not only helpful to understand the personal networks’ structure,but also can provide data support for such application,recommending friends,personalized service and information push and so on.In addition,the circles discovery of personal network for public security organs is also very important.The public security organs can study a personal network of suspects,then determine and find all the social circles accurately.It is helpful to combat criminal gangs,prevent crime,and find out other criminal suspects who are not yet found.This paper explores the problem of social circles discovery of personal network from the aspect of social networks’ communities discovery algorithm,personal networks’ circles discovery algorithm and personal networks’ circles discovery algorithm improvement.First,this paper researches the social network and the classic social networks’ communities discovery algorithm,and summarizes the ideas,advantages and disadvantages of these algorithms.Then,this paper researches the social circles discovery algorithm of personal network,DSCEN algorithm and DSCEN fast algorithm.After combining the structural information and attribute information in the personal networks,the two algorithms propose a new idea of social circles discovery.They not only can accurately divide the overlapping network in the personal networks,but also can get the features of each circle.In addition,the node similarity and node centrality can influence the social circles division of personal networks.So,this paper provides the improved scheme of DSCEN algorithm based on node similarity and the improved scheme of DSCEN fast algorithm based on node centrality.Finally,the improvement scheme of the algorithm is realized by programming,and the related experiments are carried out on the real personal networks.It is proved that the improved schemes of the algorithms can improve the accuracy of the original algorithm.