The Community Partition Study Of The Shanghai Stock Market Based On Complex Network Community Technology

With the deepening of globalization,economic competitions are increasingly fierce,so to maintain stable and healthy development of the financial market is of great significance.The stock market,as a major component of financial market,triggered more attentions in recent years.After the stock market was proved to be a complex system,many scholars turned to do researches on it by using the theory and method of complex network.Complex networks is the abstraction as well as the describtion of complex systems,of which the cells in complex systems can be abstracted as vertices and relations of cells as edges.It prominently stresses topology characteristics of the system structure.The modeling and analyzing to stock market with the full use of the complex network theory can grasp the market structure at a more macro level.Also,some of trivial details are abandoned,but we can get some valuable information which can not be obtained by other methods.Therefore,the complex networks theory provided us with a new perspective in conducting stock market research.This paper built a complex network and MST network of SSE 180 index component stocks by using complex network theory,then analyze the nature of this network to understand the law of the stock market volatility and its structure characteristics.Furthermore,based on the community division theory,divide the two networks into clusters.The main works of this paper are as follows:1.The date of stock return rate from January 1,2015 to September 30,2016 of SSE 180 index component stocks was selected to build a complex network based on the related threshold method and minimum spanning tree method.2.By calculating network metrics to judge whether or not the network have the small-world and scale-free characteristics,such as average path length,clustering coefficient,degree and degree distribution.At the same time,analyze the MST network so as to accurately find out the stock which plays an important part in the whole stock market.3.Made a progress in CNM algorithm to be employed in undirected weighted network.And with it divided stock correlation network and MST network into clusters based on community division theory,then propose a more simple and effective portfolio strategy for the majority of investors with analysis of the community structure.