An Ensemble LGBM (Light Gradient Boosting Machine) Approach for Crude Oil Price Prediction

Crude oil is considered one of the most important resources in the world today. Most of the fuel used today is refined from crude oil. Fuel also has a great impact on the global economy. The crude oil market is liquid and uncertain. The prediction of the crude oil market price has become a necessity of every second for governments, industries, and individuals. Predicting the price of crude oil can help to achieve a sustainable economy. The goal of this study is to forecast crude market prices as accurately as possible using machine learning and ensemble learning methodology. In this study, we propose the prediction of crude oil using Light Gradient Boosting (LGBM), Random Forest ensemble machine learning algorithm, Lasso Regression, and Decision Tree machine learning algorithm. The BRENT time series crude oil data are used for analysis and form a prediction model that gives less error and more accuracy. We have compared the prediction result of LBGM with Lasso Regression, Random Forest Regression, and Decision Tree regression analysis. A comparison curve is used for introducing the result, turns out LBGM gives the most accurate and efficient prediction result. We have validated our result by evaluating the root mean square error (RMSE), mean absolute percentage error (MAPE), mean squared error (MSE), mean absolute error (MAE), and the results obtained by the proposed model are significantly close and superior.