Comparison of Stock price prediticon based on XGBoost and GARCH

Abstract

For a long time, financial issues have been widely discussed by all sectors of society. There are many studies on the stock market, and the main purpose is mostly to predict stock prices and the overall trend of the stock market more efficiently. In this paper, the XGBoost Model and the GARCH Model are established in terms of the Shanghai Composite Index data. To be specific, the models are fitted and predicted by Python and Eviews, in order to find a better prediction mechanism. The XGBoost Model proposed in this paper is not satisfactory in terms of fitting and prediction effects, and has a certain degree of deviation. The GARCH Model shows a better performance in the short term. This research aims to find mathematical models that can effectively fit and predict the stock market through the combination of qualitative analysis and quantitative analysis. These results shed light on rationalizing the improvement of stock price prediction methods through simulation results.