Research Of Structure Similarity-Based Incremental Updating Algorithm On Dynamic Complex Networks Community Detection

With the development of science and technology,the change of the theory of science research, complex networks becomes a hot research topic. At present, it has found some characteristics of complex network structure such as small-world, scale-free and power-law distribution characteristic, and community structure is more important and is widely researched.Under the background of large capacity storage and the age of Big Data, at present networks demonstrate the characteristics of large scale and frequent changes. Past community detection algorithm cannot be simply transplanted from the traditional small static network directly to large-scale dynamic network platforms, because it is inefficient, and even is not practical. Therefore, according to the characteristics of the fields in the form of network,it needs to come up with a specially applied to dynamic large-scale complex network community detection algorithm, which can be used to update changes in the community caused by changes in complex networks.On the basis of the cosine similarity,this paper comes up with a structure similarity definition which is Used to describe the distance of two points:secondary impact-based structure similarity definition. What is called secondary impact-based structure similarity,is that in the network edge changes not only affect the two endpoints,but also affect the neighborhoods of the two endpoint nodes. According to the characteristics of large scale complex networks, on the basis of structure similarity definition and Related definition, this paper comes up with a static community detection algorithm which bases on local structure similarity between nodes-LBS (Local method -based on Similarity) algorithm.LBS algorithm uses the local ideology to detect community.It greatly reduces the time complexity of the algorithm. LBS algorithm adopts two stage to detect community:the first stage, through the node connection preference,it forms a small community structure; The second stage, through the role of nuclear node of connection preference chain, it connects the connection preference chains, and eventually forms the community structure. Because the LBS algorithm is advantageous for incremental updating community, on the basis of LBS,we also put forward a incremental updating algorithm for dynamic complex network community detection-IU-LBS (Incremental Updating LBS) algorithm. IU-LBS algorithm firstly detect changed node set which can impact community detection, then use the corresponding method to update the nodes of different category.Based on multiple real network and artificial generation network running on LBS static algorithm and IU-LBS dynamic algorithm,it further prove LBS algorithm for static complex network community detecting, and IU-LBS algorithm for dynamic complex network commuity updating, is efficient and accurate.