Research On Stock Selection Based On Copula Method And Clustering

After more than 20 years of continuous development of the stock market changes, and now it has become more mature and perfect. However, under the background of the trend of economic globalization and financial integration, China’s stock market is becoming more and more closely linked with the world. At the same time, as the magnitude of the stock data is growing bigger and the relationship between the financial variables is becoming increasingly complex, we have to face the following problems in research on financial risk measurement and Investment Portfolio: first, a single copula model can not meet the requirements of researching the complex relationship between stocks; second, stock researches through markets, plates or several stocks cannot meet investors’ demand for a huge investment portfolio. This essay is based on copula theory, applying three Archimedean copulas which are widely used in financial variables to form a mix-copula model. We use this model to analyze all the A shares’ data from the Shanghai stock and calculate tail dependence between stocks, then use the result to do some cluster analysis. At last we compute some charts to make our analysis much more direct and provide several effective suggestions for the investors. The empirical analysis shows that mixed copula model for stock between correlation analytical ability is better than the single copula function, and the use of cluster analysis can also help us research the difference between the stock on the upper tail(income) and lower tail(risk) dependences, also, the visual reality of tail correlation matrix for investors can provide more intuitive guidances.