Local Community Detection Based On Triangle Structure

With the scale of the real world complex network is getting larger,it is difficult to grasp the global information of the network,the complexity of the classic community detection algorithm is getting higher.Therefore,a community detection method based on local community in the network is proposed.Local community detection is a method that does not rely on the global information of complex network to detect community,but based on initial node or community,through the local information of node or edge to diffuse community.Local community detection is more suitable for the large scale of complex network than other method.The main work of the paper is as follows:1.In this paper,we find that there are two problems in the local community detection:the original position of the initial seed has an important influence on the final result of the community diffusion.After finding the initial seed,the initial Community diffusion stage is slow;2.Since the original position of the selected seed has an important effect on the final result of the community diffusion,in order to select the initial position of the initial seed more central,the community is more stable in the diffusion stage.This paper is based on the seed selection method of the core triangle,propose a local community detection method based on core triangle.The experimental results show that TLCD is superior to other algorithms in most cases for the detection of local communities;3.In order to solve the problem of the slowness of the initial community diffusion,this paper proposes a method of community detection based on multi-triangle group expansion.The algorithm chooses the initial community by searching the multi-triangle group of the core node,add missing nodes and merge redundant communities,and finally deal with overlapping nodes to get the final division of the community.In this paper,the MTCD algorithm is experimentally analyzed on the synthetic network and the real complex network.The experimental results show that this algorithm has advantages in local community detection.