This study mainly uses artificial intelligence and machine learning technology to build stock selection models to help investors choose stocks reasonably. In this paper, six machine learning models were constructed for comparison and backtesting based on the framework of the machine learning stock selection. By comparing the model classification accuracy, AUC, and other index, XGBoost and Random Forest were selected, and the portfolio was constructed. According to the analysis, the portfolio could obtain an above-average rate of return, and the portfolio obtained a net value of about 1.5 times that of the benchmark portfolio during the two-year investment test period.
展开▼
