An Empirical Study On Timing Strategies Based On Wheeling Phenomenon Of Large And Small Stock Markets

There have been many market anomalies in the domestic stock market that chould not be explained by the market efficiency hypothesis.The most representative one is the Wheeling Phenomenon of Large and Small Stock Markets.Based on the study of this periodic plate rotation phenomenon,this paper aims to establish a quantitative model to predict the time nodes of style switching between large and small markets.In recent years,with the technological innovation of artificial intelligence,many practical results have been achieved in the research of quantitative model of stock using machine learning algorithm.The research in this field has important theoretical and practical significance.As a classical machine learning algorithm,Support Vector Machine(SVM)has good performance in pattern recognition and predictive regression.By combining the advantages of SVM model and taking timing strategy based on the style rotation effect of large and small markets as the research topic,this paper aims to construct market timing strategy through in-depth study on the style rotation effect of domestic large and small markets,and use SVM model to predict the timing of style switching between large and small markets,so as to provide decision-making reference for investors.Compared with the existing research literature and methods on the effect of stock market style rotation,the innovation of this paper lies in:Firstly,this paper chooses machine learning algorithm to update the previous models.The model used by previous researchers for large-andsmall-disk style rotation research is ineffective in the current market.In this paper,the model is updated iteratively to bring new vitality and adapt to the new changes of the current stock market.Secondly,a series ofderivative factors are added to the model.Observe the change of market sentiment and market expectation from another dimension.In this paper,the design of quantitative model is systematically analyzed from five aspects: main idea,stationarity test and causality test,feature creation,optimization and optimization of kernel function and parameters,and strategy retest.The results show that the trainingcompleted timing model performs well in the sample training set and the test results are also satisfactory.Empirical tests are used to prove the effectiveness of the SVM model in the quantitative timing strategy based on the rotation effect of large and small disc styles.