MACHINE LEARNING ALGORITHMS IN FORECASTING: A COMPARATIVE ANALYSIS OF LINEAR REGRESSION, RANDOM FOREST, AND XGBOOST

Abstract

Forecasting plays a crucial role in various domains such as finance, healthcare, energy, and economics. With the advancement of machine learning techniques, predictive accuracy has significantly improved compared to traditional statistical methods. This study presents a comparative analysis of three widely used machine learning algorithms—Linear Regression, Random Forest, and XGBoost—in forecasting tasks. Using benchmark datasets and standardized evaluation metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R² score, the performance of each algorithm is analyzed. The results demonstrate that ensemble-based models outperform linear models in capturing complex nonlinear relationships, while linear regression remains effective for interpretable and low-variance datasets.