Statistical Identification And Application Of Extreme Fluctuation Events In Chinese Stock Market Based On Wavelet Denoising

In recent years,General Secretary Xi Jinping has repeatedly stressed in his important speeches on capital markets and financial work,It is necessary to follow the law of financial development,prevent and resolve financial risks,maintain financial security,deepen financial system reform,and enhance the economic ability of financial services.As an important participant in the capital market,the securities market should closely focus on the origin of the service entity economy,Through business transformation and management innovation,we will contribute to the promotion of supply-side structural reform and support for the development of strategic emerging industries,There must be new ways to achieve high quality development.These important assertions continue to show that "finance is an important core competitiveness of the country."The stock market is an important part of the securities market,and it is a complex nonlinear evolution system.Whether it is money,taxation,finance or external environment,emergencies and many other factors will have an impact on the volatility of the stock market.These influencing factors have complex coupling relationships,and the impact on market volatility has different time lag and intensity.Although China's stock market has only a history of more than 20 years,it has experienced seven surges and plunges.If we can study the characteristics of extreme volatility in the stock market,it will have theoretical reference value for strengthening market risk management,guiding investors' rational investment,and ensuring the sustainable and healthy development of the stock market.First of all,this paper introduces the relevant theory of wavelet denoising.Wavelet decomposition and reconstruction can be used to remove the interference of noise contained in the stock market sequence,and improve the accuracy and effectiveness of extreme value recognition.Subsequently,two basic characteristics of the stock market event sequence are introduced: fractal features and long-range correlation.A new method for studying multifractal features developed with fractals-Multifractal Detrended Fluctuation Analysis,combined with surrogate data(The combination of the two methods is referred to as SMF-DFA),To determine extreme thresholds for extreme events,and give the specific steps of the algorithm,identify extreme values from a new perspective.Finally,in order to verify the effectiveness of the SMF-DFA algorithm,10000 data of the X component of the Lorenz equation is used as a test set to simulate the actual chaotic time series.In the case of the known data normal range,several outliers are artificially manufactured,the SMF-DFA is used to identify the outliers,and the identified extreme thresholds are accurate and valid,which proves that the SMF-DFA method is applicable to the research in this paper.On the basis of theory,this paper selects the representative daily price data of Guizhou Moutai stock and Ziguang stock from 2007 to 2018 for empirical analysis.The original data is processed in the following order: smoothing,soft threshold db4 threelayer wavelet decomposition denoising,and SMF-DFA method to identify extreme values.Finally,the following conclusions are drawn:(1)The daily closing price data of Guizhou Maotai and Ziguang shares are non-stationary time series,and the yield data after the first-order difference are stationary time series;(2)The distribution of Guizhou Maotai yield series and Ziguang stock yield series are not uniform.The uneven distribution characteristics of the Guizhou Maotai yield series are more obvious,and the peak characteristics of the Ziguang stock yield series are more obvious;(3)The yield series of Guizhou Maotai and Ziguang have multiple fractal features,and the multi-fractal features of the Ziguang stock sequence are more intense;When the order q is fixed within a certain range,the small fluctuations and large fluctuations of the Guizhou Moutai yield series are positive and persistent,while the small fluctuations of Ziguang shares are positive and persistent,but the large fluctuations are anti-persistent;Guizhou Moutai's income is relatively stable and the risk is small,while Ziguang's earnings fluctuations are more complicated and the investment risk is greater;(4)The daily yield data of Guizhou Maotai and Ziguang shares have long-range correlation,and when q is fixed at-4,the long-range correlation index is H(q)= 0.7310 and H(q)= 0.7162.That is to say,both sequences have state persistence at this time.That is,if the sequence fluctuation is upward in a certain period of time,then the next period of time is likely to be upward;(5)The minimum threshold and the maximum threshold of the daily yield series of Guizhou Maotai are J ?-0.02766 and J ? 0.024569,and extreme fluctuations occurred 22 times,and the years of extreme events were 2007,2008,2010,2014,2015.Among them,there were 4 occurrences in 2007 and 2010,7 occurrences in 2008,6 occurrences in 2015,and 1 occurrence in 2014.In terms of volatility,the volatility of 2014 was the strongest,and the volatility was strong in 2008 and 2015.The maximum threshold and the minimum threshold of the daily yield series of Ziguang are respectively J ? 0.044031 and J ?-0.0657,And extreme events occurred 38 times,respectively in 2007,2008,2013,2014,2015,2018,the frequency of occurrence is 3 times,1 time,3 times,1 time,29 times,1 time.The highest frequency and the highest volatility are in 2015,and the Ziguang share sequence is greatly affected by the 2015 China stock market crash;(6)Wavelet denoising improves the effectiveness of the SMF-DFA method for the identification of extreme values in the stock market.