Construction of short-term stock price prediction algorithm based on MLP and CART Bagging ensemble learning

To ensure the interests of stock investors, it is essential to make short-term predictions on stock prices and reduce the investment risk of stocks. In this paper, we use Multi-Layer Perceptron (MLP) and Classification And Regression Tree (CART) algorithms and their two models based on Bagging ensemble learning (MLP-CART). Then, it discusses the impact of various factors that maaffecton stock prices on closing prices. The data of Ping An Bank are applied to the above model, and the results show that : The MLP-CART integrated algorithm constructed in this paper is applied to short-term stock price prediction, and its root means square error (RMSE) is only 0.039, 0.007, and 0.029 lower than MLP and CART. Its coefficient of determination (R2) is almost as high as 0.9, which is 0.03 and 0.26 higher than MLP and CART, respectively, indicating that its overall performance is more robust and generalizable than base learners. In summary, the research of this paper is conducive to the study of scholars ' prediction model of stock price and has played investment advice and quantitative trading advice for stock investors, which has great help to reduce investment risk.