Analysis Of Stock Market Monitoring Based On Autocorrelation Control Theory

In recent years, with the arrival of economic globalization and aggravation of the market competition, the quality has become the most powerful strategic weapon for an enterprise. The control chart is an effective tool for the quality guarantee and is widely used. The economical fluctuation is an inevitable phenomenon in the development of a country, but drastic economic fluctuations are the negative impact to people’s living. The stock market is as a barometer of the economy, which has been a key issue in the finance research. In this paper, these two hot issues are well combined, and the theory of SPC is introduced into warning and controlling of the stock market.In this paper, the basic theory of conventional control charts and the performance indexes, including the two types of errors, the average run length(ARL), the average production quantity(APQ) and the judging criterion of SPC of control charts are first introduced. Then, three frequently used types of normal control charts are also introduced, which include the Shewhart control charts, CUSUM control charts and the EWMA control charts. In this thesis, the constitutions and designing philosophies of these control charts are mainly elaborated.In order to look for effective SPC methods for autocorrelation process, first of all, the autocorrelation and time sequence theory are introduced in this dissertation. Then, the theory and designing philosophy of residuals control charts are elaborated. Finally, the process control method based on the GARCH model is provided, which is used for the clustering property of fluctuations in economic or financial data.In order to prove that Shewhart control charts become invalid for autocorrelation process, the method of Monte Carlo simulation is used. Then, effects of autocorrelation on control chart are studied. The results show that positive correlation makes the control chart too sensitive and the condition of negative correlation is just the opposite. Therefore, when the autocorrelation exists, how to choose appropriate residual control charts becomes particularly important.Finally, the 10-year data of Hushen 300 yields are used for the empirical analysis. The results show that the volatility clustering, asymmetry and the fat tail of financial time series exist. For such problems, the GARCH model is established first. Then, a series of tests on the residual control chart are performed and the results basically meet the assumptions of normal control charts. So residuals control charts are constructed, and the test results are satisfactory. At last, the out of spots are reasonably explained to achieve the effect of monitoring and early warning.