Research On The Construction Of Multi-indicator Stock Selection Intelligent Investment Strategy

With the development of the era of big data and artificial intelligence,intelligent investment has gradually replaced the artificial investment advisers financial services.Intelligent investment adviser(Robo-advisor)can also be called a robot,intelligent finance,financial adviser automation etc..Based on big data and intelligent algorithm,we build quantitative model,match investor’s risk preference and match corresponding expected revenue,and intelligently recommend stock portfolio recommendation,and implement automatic strategy transaction service.Intelligent investment can examine investors a full range of financial conditions,accurate allocation of the personal wealth,many assets interval ratio,such as stocks,funds,insurance,etc..Compared to the traditional artificial intelligent investment,investment in providing digital asset allocation for clients,can improve the accuracy of the revenue,reduce the time and labor costs,improve service efficiency.Intelligent investment expansion trend of many financial institutions around the world have a trend which cannot be halted,increase the research of intelligent investment,to become one of the hottest research topics.In this paper,aiming at the problem of the effectiveness of smart investors in the field of stock selection,we set up a multi index model strategy for the A share market,explore the characteristics of high yield stock,and establish a linear regression model(lm)strategy for risk neutral by establishing a generalized linear model(glmnet)strategy for risk averse,and establish a XGBoost Tree model strategy for the risk preference.It is concluded that the three groups can obtain the effective model strategy which is higher than the market’s excess return.Finally,the related suggestions are obtained.In the empirical part,this article data range covers all A shares.In the input variables for the first time to create four major indicators system,respectively,technical indicators,fundamental indicators,public opinion indicators and trading indicators,a total of thirty-four factors,to assess the stock’s future rate of return.In this paper,we propose to build a model pool,which includes the main regression models and classification models in machine learning.It sets up corresponding continuous and discrete evaluation indicators for different models,assists in screening the optimal model,then adjusts the parameters of the optimal model,and constructs the output variables as The difference between the model’s stock picking yield and the broader market’s yield;in the strategy evaluation part,all strategies are evaluated and ranked using Sharp’s ratio,and three types of groups are constructed,namely risk averse,risk neutral and risk preference.Customers with different benefits and risk requirements match different investment strategies.The results of this study and corresponding numerical calculation procedures can provide support for related enterprises and investors in intelligent stock selection.Finally,this paper summarizes the contribution of stock-picking strategy to the investment advising and proposes the follow-up research direction for the shortcoming of this article.At the same time,it also looks forward to the future development of the smart investment and puts forward some legal restraints.