Research On Robo-Advisor Strategy Based On Overall Performance Prediction Of Public Funds

The field of technology finance is a frequently occurring field of technological innovation and method innovation.With the development of big data technology and the improvement of computer computing capabilities,artificial intelligence is increasingly widely used in the field of technology finance.As a product of the combination of artificial intelligence and technology finance,Robo-advisor has become one of the hot spots in the field of Internet finance.The theoretical framework of Robo-advisor includes modules such as user portrait,asset prediction,asset evaluation,asset selection,asset allocation,and holding strategy.It is a business model that systematically analyzes investor characteristics and provides investment suggestions.This dissertation focuses on the empirical performance of Robo-advisor strategies in the fund market,using funds as investment targets.It mainly constructs an Robo-advisor strategy that includes fund market trend prediction,fund performance prediction,and portfolio design,applying this strategy to the fund market.This dissertation forecasts the SSE FUND INDEX to determine the future trend of China’s public fund market.Customizing equity fund,establishing multiple machine learning models,building indicator systems from fund performance indicators,fund manager indicators,and market trend indicators,this paper predict fund returns,and calculate corresponding risk indicators.Designing portfolio strategies corresponding to different risk preferences,this paper apply them to data sets,and dissertation the empirical performance of Robo-advisor strategies.Empirical evidence shows that the Robo-advisor strategy constructed in this article can earn excess returns for investors.Considering transaction costs,the intelligent investment advisory strategy in this paper can still effectively adjust risks and returns.In this paper,it was found that most machine learning algorithms(LASSO,Ridge,Elastic Net,and SVM)performed better than benchmark models(OLS)in predicting fund returns,while three nonlinear algorithms(GDBT,XGBoost,and DFN)performed best in empirical terms.