Title: Forecasting the stock direction using ensemble random forest

Authors: K. Senthamarai Kannan; V. Parimyndhan

Addresses: Department of Statistics, Manonmaniam Sundaranar University, Tirunelveli – 627 012, India ' Department of Statistics, Manonmaniam Sundaranar University, Tirunelveli – 627 012, India

Abstract: Predicting stock prices is a challenging task due to market fluctuations and the nonlinear nature of stock markets. Forecasting errors can increase investment risk, so reducing them is essential. In this study, financial data including high, low, open, and close prices of stocks were used to generate new variables. Here TCS stock data is employed to construct a Random forest model. This model predicts whether a stock will close positively or negatively based on past values of technical indicators. The findings in this work demonstrate that, using random forest for stock forecasting is beneficial for both institutional and retail stock investors. The unique contribution of our work is the utilised of technical indicators as high-accuracy features for predicting the direction of stock prices. The performance accuracy rate of 74% for the model is obtained by increasing the number of training cycles for the algorithm.

Keywords: forecasting; random forest; RF; technical indicators; indicators; machine learning; ML; stock.

DOI: 10.1504/IJMOR.2024.140069

International Journal of Mathematics in Operational Research, 2024 Vol.28 No.3, pp.396 - 407

Received: 15 Mar 2023
Accepted: 02 Apr 2023
Published online: 19 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article