Design And Application Of Quantitative Stock Selection Model Based On Artificial Intelligence

With the development of artificial intelligence,quantitative stock selection models based on artificial intelligence are born.Most of them adopt intelligent optimization algorithms or machine learning prediction methods to realize intelligent stock selection.These quantitative stock selection models have achieved excellent results in capital markets in both developed and developing countries.This paper also uses intelligent optimization algorithms and machine learning prediction methods to construct quantitative stock selection model.The innovation of this paper is to construct a new stock selection model by combining stock forecasting with stock scoring for the first time,which considers not only market-based forecasting factors,but also the fundamental factors of enterprises.The model consists of two steps,stock prediction and stock scoring.Firstly,the machine learning prediction method is used to predict the future price of each stock,and then the forecasting factors of the stock are constructed through the future price.Secondly,the forecasting factor and various fundamental factors are introduced into a multi-factor stock selection model to evaluate the value of each candidate stock and select high-scoring stocks to generate an equal weight portfolio.This paper applies this model to the empirical study of A-share market,and makes a systematic comparison with other benchmark models(different prediction models,different factor designs,different optimization algorithms,different fitness functions).Empirical results show that the quantitative stock selection model proposed in this paper has achieved remarkable results in the A-share market,and the portfolio returns are much higher than the average market performance(i.e.the equal weight portfolio of all candidate stocks)and China's A-share index.It is worth noting that the forecasting factor constructed in this paper is given a high weight in the model optimization.This discovery shows that the forecasting factors constructed in this paper play an important role in stock selection.It also supports the new idea that introducing stock forecasting factors can improve stock selection decisions.