Forecast Of Stock Price Fluctuation Trend Based On Move Trendency Entropy Dimension

As an important part of the financial market,the stock market plays an important role in providing direct financing channels and allocating financial resources effectively.Whether the stock price fluctuation trend can be accurately predicted is directly related to the effectiveness of stock market risk management.It has a very clear practical significance for financial regulators to take regulatory measures in advance to prevent abnormal fluctuations in stock prices.With the increasing complexity of the financial market,the stock market shows more and more fractal non-linear characteristics.However,the traditional linear regression analysis method obviously can not describe these complex characteristics,and can not be directly applied to the prediction of stock price fluctuation trend.In this paper,the representative leading stock in China's stock market is taken as the research object,and the following research is carried out around the topic of non-linear prediction of stock price fluctuation trend,and the corresponding conclusions are obtained.(1)Using R/S,DFA and MF-DFA methods to characterize the fractal characteristics of stock samples,the results show that R/S,DFA and MF-DFA methods all show the fact that China's stock market has fractal characteristics.By comparing various characterization methods,we can find that R/S has limitations in describing the fractal characteristics of stock market because stock samples do not satisfy the assumption of time series stationarity;DFA can only describe single fractal characteristics,but also not suitable for describing complex fractal characteristics of stock market;MF-DFA can show the stock market as a whole.The fractal characteristics of the stock market can also excavate the local situation of the fractal characteristics of the stock market,so that the fractal characteristics of the stock market can be described more comprehensively and completely.(2)The prediction results of moving trend entropy dimension and entropy dimension are compared.The results show that the entropy dimension can only depict the persistent effect or reversal effect,while the moving trend entropy dimension can not only depict the persistent effect or reversal effect,but also divide the rising and falling states,and calculate the entropy dimension in both states,so as to investigate the persistent effect and reversal in both states.As a result,the advantage of moving trend entropy dimension is more obvious than that of entropy dimension.(3)The prediction accuracy is compared between the moving trend entropy dimension and CAPM regression model.The results show that the prediction accuracy of each stock sample using the moving trend entropy dimension is higher than that using CAPM regression model,which indicates that the moving trend entropy dimension model can capture the stock market.Fractal non-linear characteristics and effective prediction by judging the direction of fluctuation trend,so it has better prediction performance than CAPM regression model.At the same time,the moving time limit of the moving trend entropy dimension method can generally obtain better prediction results in 60 days than in other periods,which indicates that the setting of the moving time limit is neither the minimum better nor the maximum better,but needs to set the suitable moving time limit according to different financial assets.(4)The performance of the mobile trend entropy dimension and CAPM regression model is compared from the perspective of investment strategy performance.The results show that,compared with the traditional CAPM regression model,the moving trend entropy dimension model has the highest cumulative yield and information ratio.It shows that for the actual investment,the use of trend entropy dimension method can effectively realize the investment strategy of selling high and buying low,and play an important guiding role in the promotion of investment returns and the avoidance of investment risks.At the same time,the cumulative rate of return and information ratio of the moving trend entropy dimension model are the highest under the 60-day limit,which shows that the best prediction effect can be obtained by setting the moving period at 60-day,which also has a very important reference value for investment practice.(5)The robustness of the prediction performance of the moving trend entropy dimension is tested.The results show that the moving trend entropy dimension model still has better prediction results than the entropy dimension model and the traditional CAPM regression model,which proves that the moving trend entropy dimension model has good prediction performance and sufficient robustness.Based on the above conclusions,the moving trendency entropy dimension model proposed in this paper has a good effect in forecasting the fluctuation trend of stock price.It is superior to the entropy dimension model which can not judge the direction of fluctuation and the CAPM model which can not describe the non-linear fractal characteristics,so as to carry out relevant investment and wind for financial subjects.Risk management financial activities provide targeted application tools and methods.The financial supervision department can use this model to judge the fluctuation of the stock market,so as to take measures to eliminate the hidden risks ahead of time,so as to maintain the stability of the stock market;and for investors,it can use this model to judge the trend of the stock market,by adopting "buy low and sell high".Investment strategy to obtain excess investment income.