Research On The Discovery Method Of Community In Complex Network

With the rapid development of modern science and technology, in real life the scale of different networks in different fields is becoming larger and larger. The amount of information it contains is also complicated. With the increasing level of human technology and the deepening of understanding, scholars have found that the complex systems of many different industries in different areas present the common nature, which are complex networks. In addition to the small average path of the statistical properties of small world, the preferences of node connections, power law degree distribution of pay scale free distribution of statistical properties, community structure also constitutes an important complex network structure. The community structure means the close-knit of nodes within the community and the sparse regiment between communities. The discovery of community structure in complex network is of great theoretical significance for people to analyze the topological structure in complex network, understand its hidden pattern and predict the behavior of complex network. Also, it has a wide range of applications in social communication networks, biological networks and the Internet.This paper makes the relevant research on the discovery algorithm of community in complex network. The paper first introduces the research purpose and significance of community structure in complex network and related challenges. Next, make a simple description of the basic knowledge of complex network and their representation. Then expound the traditional discovery method of complex network by graph partitioning and hierarchical clustering. The classical graph algorithm includes Kernighan-Lin algorithm and the spectral bisection method. The hierarchical clustering algorithm includes GN based on splitting and NF based on cohesion. Then describe the core idea of these algorithms and the inadequacies of the current algorithm.Based on the analysis of traditional algorithm, this paper adopts the thinking of hierarchical cohesion and puts forward the algorithm to find the community existing in complex community by the optimization of modularity. Also give a detailed description of the improved algorithms proposed in this paper. Use small network and vividly describe the specific implementation processes of the proposed algorithms by the network's visualized evolution. Finally, by the standard data and actual data, the community structure in baseline network discovered by this algorithm is very close to the actual situation. There is a certain credibility to find the community structure in actual network, which verifies the reliability and efficiency of the proposed algorithm.