A Comparative Study Of Quantitative Stock Options

In recent years,machine learning has been gradually applied in various fields,such as medical treatment?transportation?face recognition,In addition,in the financial field,machine learning is also trying to use as an auxiliary investment,As an important investment target of quantitative investment,the use of machine learning method to explore the future trend of stock is more and more concerned.In this paper,the quantitative stock selection scheme is designed from two aspects: the traditional quantitative stock selection scheme and the quantitative stock selection scheme based on machine learning.In the traditional quantitative stock selection scheme,based on CAPM model,Fama-French three factor model and Fama-French five factor model,this paper designs three kinds of quantitative stock selection schemes,and uses the data of July,August and September 2019 to simulate transactions.The results show that when the market is in a bull market,quantitative stock selection scheme based on CAPM model has the highest return and the minimum max drawdown;When the market drops slightly,among the quantitative stock selection schemes with positive returns,the quantitative stock selection scheme based on Fama-French three factor model has the highest returns,while the quantitative stock selection scheme based on CAPM model has the smallest max drawdown;When the market falls seriously,the decline range of the above three quantitative stock selection schemes is higher than that of the CSI 300 index.In the quantitative stock selection scheme based on machine learning,firstly,the factors are determined from the basic and technical aspects,In this paper,119 candidate factors are selected,According to the Spearman correlation coefficient and its significance,the factor validity is tested with the data of June,July and August 2019 respectively,Then,based on random forest,neural network and xgboost,three quantitative stock selection schemes are designed.The results show that when the market is in a bull market,the quantitative stock selection scheme based on xgboost model has the strongest profitability and the quantitative stock selection scheme based on neural network model has the smallest max drawdown;When the market drops slightly,the quantitative stock selection scheme based on the neural network model has the highest return and the minimum max drawdown;When the market falls seriously,all schemes fall,but the quantitative stock selection scheme based on the neural network model has the lowest drop.In addition,this paper proves that the factor has timeliness.Finally,compared with the traditional quantitative stock selection scheme and the quantitative stock selection scheme based on machine learning,when the market rises or falls sharply,the quantitative stock selection scheme based on machine learning performs better,When the market falls slightly,the traditional quantitative stock selection scheme and the quantitative stock selection scheme based on machine learning have their own advantages and disadvantages.