Research Of Structure In Stock Markets Based On Spectral Clustering

With the continuous deepening of economic reform, China’s stock market is gradually improving and developing, the stock market has become the important component of China’s securities industry and even the whole financial industry. Listed companies is rapidly increasing, it is not easy pestered beyond endurance for investors to understand the stock one by one. In order to adapt to the situation and needs, some financial services organizations used the advantages of their business knowledge and the familiarity with market, compiled for the stock price index. And the stock price index can’t fully reflect the market, Therefore, this paper through the analysis of plates, try to get more comprehensive understanding of the stock market from the macro point of view.In recent years, with the development of data mining technology, using the data mining technology has become a hot topic. In the plate study, mostly using the traditional clustering methods, in this paper, We transfer the Close Price from December2008to March2012of the33Plate index, using the spectral clustering algorithm to cluster33plates, calculating for each class of plate correlation, comparing of each class of yield and growth. The paper uses the K-nearest neighbor and then selects1neighbor of each section, constructing a1neighbor network. And then calculate the betweenness centrality of the network, research network topology. Against the angle of contact to examine the stock, rather than isolated point of view to examine the relationship between the stock reveals the structure of the stock market, and thus we can have a broad understanding on the stock market.In this paper, the spectral clustering method and betweenness centrality of the network were used to analysis the plate index. The results show that, through the understanding of several core plates, we can understand the stock market. Research in this period of time, the Shanghai Composite Index, textiles, garments, engineering and construction, electronic information, commercial chains, the Shenzhen Component Index is a core plate of each cluster, they are representatives of the others in the same clusters, we can understand a cluster of plates by understanding a certain type of core plate. The results of these studies plays a useful role in helping investors understand the overall stock market structure, and helping investors to grasp the entire stock from a macro better.