Comparison Of Linear And Nonlinear Multi-Factor Stock Selection Models In A-Share Market

In recent years,quantitative investment has mushroomed in China.The rise of quantitative investment will lead the A-share market to a professional and mature direction.At present,this new type of investment method is gradually accepted by a large number of institutional investors in the market,and they have done a lot of research on it from various angles.Multi-factor stock selection model is actually the most important model of the whole quantitative model in the field of portfolio investment.From 1950 s to 1960 s in the US market,the multi-factor stock selection model has been successfully applied in overseas markets.However,in our country,multi-factor model is also the most widely studied and applied model in the field of quantitative investment.Previous studies have focused on the selection of validity factors or the improvement of the accuracy of the model algorithm itself.Few studies have focused on the comparison of the advantages and disadvantages of linear and non-linear models and their applicability in A-share market.This paper focuses on the comparative study of multi-angle linear and non-linear multi-factor stock selection models in A-share market.Combining with the actual situation of the development of A-share market and its own particularity,the requirements and evaluation criteria for stock selection in A-share market are determined.The validity of the linear and non-linear multi-factor stock selection models is tested by selecting 27 basic factors and grouping test.On this basis,we select the sample stocks of Shanghai Stock Exchange Index from January 1,2014 to May 1,2019 as an empirical test,and compare the linear and non-linear models from three perspectives: the length of different training sets,the number of positions and the switching of different market styles,so as to analyze the applicability of different models in A-share market.