Sepctral Clustering Research And Application On Community Detection Of Social Media

In our real world there are a lot of complex networks,the nodes represent the research object,and there are intricate relations between these nodes.As weibo,Twitter,social media platform in the social network is a kind of complex network structure in real society,social network has various common characteristic of complex networks,including the small-world effect and scale-free feature,group structure and so on.In this paper,we study the main object is the network structure of the group.The researches of community detection has important theoretical significance and practical application,and these issues has received extensive attention of researchers in various fields recent years.In recent years,a large number of social media have made the scale of the social network bigger and bigger.In the network,the relations become increasingly complex,and this also brings challenges to community detection algorithms.There are many classic community algorithms now such as Fast-Newman algorithm,GN algorithm,Kernighan-Lin algorithm,LFK algorithm,CPM algorithm and so on.In this paper,the related research on spectral clustering algorithm is mentioned.The specific work that has been done in this paper is as follows:1)Due to the traditional NJW spectral clustering is strongly affected by the scale parameter,so the study of this problem this paper reproduction clustering algorithm based on the introduction of the similarity measure based on gravity and combined with Floyd-Warshall algorithm,eliminate the influence of the scale parameter,and also achieved a better clustering result.2)The traditional spectral clustering uses K-means algorithm as clustering operation to feature vectors,so the choice of initial clustering center is sensitive.in allusion to this problem,this paper puts forward a new initial clustering center selection method,and then improved the stability of spectral clustering algorithm.3)This article will put forward the improved algorithm applied to the community detection and comparative experiments on three real network,and then test the effect of the community detection algorithm.4)In allusion to Sina Weibo in social media,in this paper,based on the improved community detection algorithm,we designed a community detection framework to discover communities for weibo users,and then get the further validation of the research in this paper.