Empirical Research On Quantitative Stock Selection Based On Logistic Regression

Quantitative trading is the current mainstream investment mode.Among the many stock selection models,the multiple factor stock selection model is undoubtedly the most popular one.The multi-factor stock selection model chooses the high quality stocks through scoring method or regression method to construct the portfolio by selecting effective factors.At present,the research on the multi-factor stock selection model is mai.nly based on the scoring method.But the scoring method is influenced by subjective factors.In this paper,we try to choose high quality stock by regression method to build stock Pool A.The empirical results show that the regression method of multiple factor stock selection model obtains higher excess income.On this basis,the logistic model is used to optimize the multiple-factor stock selection model.Logistic model is a kind of generalized linear model,which is widely used in the fields of medicine and credit rating.Through empirical analysis,it is proved that logistic model is an ideal prediction model.On the basis of the regression method,a multiple factor stock selection model,the logistic model is used to forecast the Shanghai and Shenzhen 300 shares,to select the stocks that can outperform,to form the stock Pool B,to compare with the stock Pool A,and to select the overlapping stocks to construct a new portfolio.The empirical results show that the improved multiple-factor stock selection model with logistic model obtains good excess income.The research significance of this paper has two aspects,on the one hand,through the improved multiple-factor stock selection model in the stock market to find high-quality stocks to build investment portfolio,and short of the Shanghai and Shenzhen 300 stock index futures,so as to achieve the effect of hedging,to obtain more stable returns.On the other hand,the LOGISITC model is used to improve the multiple-factor stock-picking model,and the result is better,so that the follow-up researchers can play a role in the study.