Research On Multi-factor Quantitative Investment Model Based On Elastic NET

With the rapid improvement of computer technology and the continuous development and reform of China's capital market,people's investment concept is more and more rational,quantitative investment based on computer technology arises at the historic moment.Quantitative investment refers to the establishment of investment model through quantitative methods,in order to obtain stable or even as much income as possible.Quantitative investment has been recognized by the majority of researchers and investors because of its scientific and effective way of operation..Multi-factor stock selection model is a very important way to quantify the stock selection species to establish a portfolio.From Markowitz's portfolio theory to the capital asset pricing model,the market excess rate of return is used to explain the excess return of individual stocks,and then the arbitrage pricing model is developed to show that the return of individual stocks is affected by multiple factors.Until on this basis,the establishment of Fama-French three-factor model,five-factor model,one after another of the theoretical results of scholars to promote the birth and continuous development of quantitative factor stock selection..The idea of quantitative factor stock selection is to describe the characteristics of the return and risk of assets through one or more indicators.In this paper,the multi-factor strategy is analyzed,and a total of 30 factors are selected from the aspects of estimation factor,growth factor,quality factor and technical factor.The empirical part of this paper starts with the regression method,using the corresponding value of the last trading day of each month of the HS300 index stocks,and selecting the data period from June 2011 to June 2018.During this period,the data from June 1,2016 to June 1,2018 are used to test back,and then summed up as the score of each stock according to the product of the corresponding factor value and the regression coefficient,and then sorted.Select the top 20 stocks to form a portfolio.In this paper,OLS regression and Elastic Net regression are used to compare and study the profitability and stability of the selected factors.In order to establish a multi-factor model to obtain more Alpha excess returns under certain risks.The results show that using the selected 30 factors,the two regression methods can obtain excess returns,but compared with OLS strategy,the Elastic Net strategy has more advantages in profitability and stability.