Short-Term Stock Price Prediction Algorithm Construction Based on Integrated Learning of SVR and RF with Bagging

Abstract

In the context of the emergence of artificial intelligence machine learning algorithms, how to handle training data for reasonable and accurate prediction of the stock future market is expected to bring effective methods, but previous traditional models are difficult to be used effectively in complex stock markets. In this paper, by exploring new stock forecasting methods such as Support Vector Regression (SVR), Random Forest (RF), and integration based on integrated learning, the model results are compared with previous traditional model results and the models are evaluated using R2 and MSE metrics. The algorithm based on Bagging integration has better robustness and generalization, in which both R2 and MSE have some improvement compared with those before integration. The research in this paper is beneficial to provide a reasonable prediction approach for stock forecasting later, which can help consumers make better quantitative trading.