The Approaches to Predict Stock Prices by Random Forest

Abstract

Abstract. People are always curious about how to predict the stock prices, various type of prediction tools are introduced to help people explore the trend of the stock price movement. With the development of the machine learning, there are more and more models and methods that we can use to predict the stock prices, while Random Forest, a typical machine learning algorithm, will be discussed in this article to examine its performance in the stock market. The data utilized in this study is sourced from Yahoo Finance, with several additional indicators computed and incorporated into feature set of the Random Forest model. To evaluate its performance, a comparison between the Random Forest model and the simple linear regression model is going to be conducted, which is a considerably basic model and serves as a benchmark for the minimum returns individuals can expect. The RMSE and MAPE will be compared firstly, which can help us assess which model demonstrates superior predictive accuracy, while the ROI derived from the strategy based on these models predictions will serve as a measure of their practical performance in the stock market. Collectively, this data will substantiate that Random Forest is indeed an effective model that is capable of assisting individuals in achieving greater profits.