Detection Of Weibo Overlapping Community Based On Multidimensional Information And Edge Distance Matrix

One of the hot topics of weibo network research is to find the users collections having the same interest preferences or frequent communication from the huge weibo network.And it divides the user set into reasonable groups.Its purpose is to achieve the precise marketing push of product or service,commodity recommendation and public opinion analysis.However,community discovery is an effective method to find the above set.The traditional community discovery algorithm is often based the network structure,without considering the unique characteristics of weibo network,such as user tags,forwarding,praise,comment and other multidimensional information,which results in the inaccuracy of community discovery.In addition,in weibo networks the users usually have multiple interests,such as users who are interested in music are also interested in dance,which leads to a result that the same node may belong to different communities.Therefore,the traditional non-overlapping community discovery algorithm can't actually depict the microscopic features of weibo social networks.In this view,this paper takes weibo networks as research object and proposes a weibo overlapping community discovery algorithm,called MIEDM.It is based on multidimensional information and edge distance matrix.First of all,this paper establishes the weighted network topology graph fusing the weibo multidimensional information such as weibo user relationship,user behavior,weibo content,and geographic location.Then,based on the edge-node-edge random walk model,this paper constructs the edge distance matrix from the perspective of the edge of the network topology graph.The matrix not only considers the distance of adjacent edges but also considers the distance of non-adjacent edges.Secondly,this paper improves the existing density peak clustering algorithm making the way of manually selecting clustering center become automated confirming by programs,and uses the improved algorithm to carry out the initial community division of the edge distance matrix.In addition,the initial discovered community is merged according to the expansion module degree,which can reduce the overlap degree among the communities and improve the quality of the community division.Finally,the validity and generality of the algorithm are verified by the weibo network data set and the classical real network data set.The results of experiments show that this algorithm has higher accuracy,stability and generality compared with other community discovery algorithms.