The Analysis Of The Long-term Memory And The Trend Of China Stock Market

Stock market of new China, since the establishment of Shenzhen Stock Exchange on December 1st, 1990 and the establishment of Shanghai Stock Exchange on December 19, 1990, has witnessed its rapid development. As of December 31, 2015, the total number of listed companies in Shanghai and Shenzhen was 2,827, the total share capital of was 43,014.82 million shares, the total market value of was 531,304.20 yuan, accounting for 83.52 percent of GDP. Classical portfolio theory to the efficient market hypothesis, produced in the 1950 s by Markowitz, William Sharpe, Fama, based on rational person, described a situatiion in the stock market, assuming that investors invested in stock market and found investment portfolios with the use of historical information, public information and insider information. Although its position as an important financial market theory can not be shaken, efficient market theory's stringent assumptions lead to the applicability be a great challenge, but it also encouraged people to turn to find a more effective interpretation theory.Fractal market theory, produced in the 1990 s, is closer to the real situation with their assumptions and a series of feasibility studies in quantitative analysis model for financial market analysis, which has a very strong applicability. This paper, based on the comparative analysis of the applicability of various quantitative model ?practicality and the consideration of the author's knowledge level, use JB test, Q statistic, BDS test to verify the nonlinear characteristic of China's stock market; with the use of Chinese stock market daily data, weekly data monthly data, this paper tries to verify the statistics self-similarity of Chinese stock market; with th use the classic R / S method, modified R / S method, DFA method, this paper tries to verify the long-term memory of China's stock market; by using the classic R / S method, modified R / S method statistics obtained V, this paper concludes the number of days of non-cycle is [270,330] of China stock market. The paper has validated the vision characteristics, inproving that China stock market does not meet the certification requirements of finance in the traditional linear paradigm, normal distribution, random walk, efficient market, etc. This shows that the use of certain quantitative analysis method can catch the trend running characteristics of Chinese stock market.Through the use of the correlation matrix, mean, standard deviation, variable coefficient, comparing with different length of the window, this paper concluded that time-varying Hurst exponent selection model of Shanghai stock index market can be useful on the time window of N=270 at R / S method and N=270 in Shenzhen Stock index market. Similarly, under the DFA method, when the Shanghai index selection model uses N = 255 time window under varying Hurst index; Shenzhen Stock Index market timing model uses N = 255 time window under varying H index. Using the R / S method comparative analysis, when DFA method under varying Hurst index selection model, the overall accuracy of the effects from market timing model point of view, the use of DFA methods derived variable H Index for China's stock market trend analysis more accurate because the variable H index next DFA method for 2008, 2015, generally a large market can give a more accurate trend analysis, but using R / S method under varying H Shenzhen Component index on the interpretation of the 2008 Great Quotes and applicability of a certain biased questions. But it can not completely abandon the use of time-varying Hurst index on R / S method, after all, R / S method on the whole still has its explanatory and applicability on the whole.