Application Of Quantitative Strategy Based On Dynamic Mode Decomposition In Chinese Stock Market

With the booming development of quantitative transactions in China,the study of methods of quantity has entered a peak.This paper has done a deep research on the quantitative mode of dynamic mode decomposition,DMD.In this article,what we discussed is mainly about: First,how to analyze the data of stock price fluctuations effectively and search patterns from it;Second,how to use the DMD model to predict stock prices through existing data,and then build a stock selection plan based on it to gain profit.Since the stock price changes often are randomness,time-series,and tendency of volatility,this article selects the Chinese stock market(excluding ST shares & suspension shares)during the period from May 6,2015 to August 3,2015 as the research object,and constructs a data matrix with two dimensions of time and daily closing price,the method of singular value decomposition is used to reduce the stock data dimension,find eigenvalues and feature vectors,and combine DMD to forecast stock prices.It is found that the stock price also has a strong trend during the period of violent market fluctuations.By observing the distribution characteristics and am-amplitude map of the Ritz eigenvalues of the DMD algorithm to compare the volatility of single stocks and the overall market and combine the existing market conditions to predict the volatility trend of the stock.Finally,comparing the stock selection strategy based on the DMD algorithm with the average yield of the Shanghai Stock Exchange Index,the DMD stock selection strategy obtains a higher yield,further confirming the effectiveness of the algorithm in the Chinese stock market.Combined with the above analysis,this paper proposes some reference suggestions such as improving the DMD quantization model,identifying the best time to use the model,and optimizing the stock selection strategy.