Research On The Stock Selection Based On Multiple-factor Quantitative Model

Quantitative investment is a kind of method based on data.It takes the mathematical model as the core,and conducts investment with the help of computer programs.Among these mathematical models,the multi-factor model is the most popular.This paper uses sorting scoring method to construct three-factor model,found that the domestic stock market has obvious small cap effect.Asset pricing is the core issue in the financial field.The difference between the transaction price and the intrinsic value of the asset determines the size of the profit.Profits come from undertaking risks.Multi-factor model is one of the most important methods to characterize these risks.In order to find out the source of risk,this paper extracts the former revaluation price data and the relevant factor data of the constituent stocks of CSI 800 from Jan.1,2012 to Dec.31,2015.Then,we use the single factor model to make back-testing directly.After that,we get the validity of the factors through comparing the risk and return indicators,such as alpha,Sharpe ratio,information ratio,and max drawdown.The most effective factor among the five factors is the logarithmic total market value,and the second is the net profit growth and the revenue growth rate,which means these three factors,can characterize risk effectively.After obtaining the effective factors,we do three more steps.Firstly,we sort the constituent stocks according to logarithmic total market value,and select the 50 stocks with the smallest market value.Secondly,we give net profit growth and the revenue growth rate the same weight to score the 50 stocks,and selected 15 stocks with the highest scores.Finally,we establish positions by the same weight,thus multi-factor model is completed.Using this model,we select new stock portfolio from the constituent stocks of CSI800 and observe its risk-return index.We found,compared with single factor model,three-factor model shows better results.