Research On Local Community Detection Algorithm

Community is the "groupuscule" in the social network.The nodes in the same community are tightly connected,while the nodes in different communities are sparsely connected.Most community detection methods require the global information(e.g.the numbers of nodes and edges in the network as well as connections between nodes)of the original network to be available,however it is often expensive(even no way)to obtain the global information of the network in many real-world networks.On the other hand,sometimes people only care about the local community to which the given starting node belongs.Therefore,local community detection has drawn wide attention.In addition,the local communities with different scales are often needed.In fact,a node often belongs to small communities and large communities at the same time.So it is meaningful to detect the multiscale local communities.This paper studies local community detection and multiscale local community de-tection,and mainly contains the following two aspects.(1)We propose two local community detection methods DMF_M and DMF_R.Taking DMF_M as an example,firstly,we analyze the problems of the existing local community detection algorithm(M method)and the characteristics of dynamic changes in the formation of local community.Then we design different dynamic membership functions for local community detection according to the characteristics of each stage.Similarly,we propose another local community detection algorithm DMF_R.Finally,we tested our method on synthetic datasets and real datasets.The experimental results show that the local communities discovered by our methods are closer to the real local communities.(2)We propose a multiscale local community detection method.Firstly,we pro-pose a local modularity LQ for measuring the quality of the local community,and we prove that maximizing the local modularity LQ is consistent with maximizing the ex-isting local modularity M,maximizing LQ and maximizing the other existing local modularity R are consistent under certain conditions.We proposed the multiscale local community detection method based on the local modularity LQ.We test our method on synthetic datasets and real datasets.Experimental results show that,for a given starting node,our algorithm can find meaningful multiscale local communities.This paper focuses on local community detection,i.e.detecting the local commu-nity based on the local information of the network(not the global information).We have tested our algorithms on several synthetic and real datasets,the experimental re-sults show the effectiveness of our algorithms.This paper has some reference value for the research of local community detection and multiscale local community detection in social networks.