A machine learning model using multi-feed data forecasts crude oil prices, with Random Forest showing the best accuracy.
Authors
Sruthi Paul, National Institute of Technology, Dept. of Computer Science and Engineering, Calicut, India
Anu Mary Chaco, National Institute of Technology, Dept. of Computer Science and Engineering, Calicut, India
Biplab Bhattacharjee, Associate Professor, Jindal Global Business School, O.P. Jindal Global University, Sonipat, Haryana, India
Summary
Crude oil price forecasting has a significant role in advancing the Sustainable Development Goals by enabling more informed decision-making in energy management and resource allocation. The accurate forecasting of crude oil prices is always a challenging task due to the inherently volatile nature of the price data. Additionally, the non-linearity and non-stationarity of the data make the price forecasting further complicated. In such cases, machine learning models can outperform the traditional statistical models. The machine learning models have the potential to analyse complex, non-linear, and non-stationary time series data and generate more precise forecasting results. In this work, we implemented the machine learning-based crude oil price forecasting model using multi-feed data. Along with the historic crude oil price data, we used financial markets’ data and Google Trends data, which give additional information about the price fluctuations. We can analyse user sentiment information using Google Trends data, which also influences crude oil prices. We developed three forecasting models using Random Forest, XGBoost, and Linear Regression models. The comparative study reveals that the Random Forest model outperforms the other two models in terms of accuracy.
Published in: ETIS International Conference on Emerging Technologies for Intelligent Systems, ETIS 2025
To read the full article, please click here.
