Based On Bayesian Quantile Regression Empirical Analysis On The Risk Of Stock Market

With the rapid development of economy, the global financial system is in constant change, especially the expansion of stock market in China, so that the dust. Because of Chinese stock market started late, in the face of such a large market risk, the development level is still in the primary stage, is not mature. The risk exists in the stock market has become the broad masses of the people, investment and financial institutions is the most concerned problem. The subprime mortgage crisis in the America, issue share measure available in the market is more prominent, in twentieth Century the tacit understanding, a new concept namely the risk value into the financial risk management (VaR), the first measure in promoting scientific above has the very good effect.Robust least square method commonly used is not high, and no longer belongs to the unbiased estimation, so if the data to be processed appeared peak thick tail characteristics and obvious situation heteroscedasticity cannot use traditional method to measure the risk value. So, if we want to avoid the defects of least square variance of quantile regression, the author put forward the theory of an asymmetric LaPlace distribution and Bayesian method combined with measures based on risk together, for the stock market prediction research has very good effect. In the following discussion, will measure theory and Bias risk in detail in the stock market the quantile regression theory. Then on the basis of these, calculation and evaluation method of risk value induction. And construct a model of a return to risk measurement, by using the inverse jump Markov Monte Carlo algorithm to complete the identification of them work, and the Bias parameters are estimated, in order to achieve the verification of the stock market risk measure method. Facts prove that the model can accurately evaluate the risk value exists in the stock market.