USING MACHINE LEARNING TO FORECAST NIGERIA'S CRUDE OIL PRICES FOR BUDGETING: A COMPARATIVE ANALYSIS

Abstract

The use of machine learning models to predict Nigeria's crude oil price—a crucial component of organizational and governmental budgeting is examined in this paper. A number of machine learning methods, including Random Forest, Extra Trees, Decision Trees, and Support Vector Machines (SVM), were assessed using historical monthly crude oil price data that was acquired from the Central Bank of Nigeria (CBN). With test RMSE = 7.7890 and MAE = 5.3054, the Random Forest model had the highest test score (R2 = 0.8921). Model consistency was evaluated using standard deviation, RMSE, and cross-validation (CV). Discussions were held regarding residuals, error behavior, and the modeling's budgeting implications. Future work with hybrid and deep learning models was outlined, along with suggested improvements. The approach and results may help Nigerian authorities provide more accurate crude price projections for budgetary planning.