Prediksi GDP dengan RF dan XGBoost Berdasarkan Aspek Sosial, Ekonomi, dan Lingkungan

Mubarok, M. Husni
Septian, Firza

MDP Student Conference
Vol 4 No 1 (2025): The 4th MDP Student Conference 2025

This study aims to analyze Gross Domestic Product (GDP) prediction using Random Forest and XGBoost algorithms by considering social, economic, and environmental variables. The dataset was obtained from Kaggle and includes 22 independent variables influencing GDP. The model was developed with Whale Optimization Algorithm (WOA) optimization to improve prediction accuracy. Experiments were conducted on the Google Colab platform, and evaluation metrics included Mean Absolute Error (MAE), Mean Squared Error (MSE), and R-Squared (R²). The results show that XGBoost with WOA optimization achieves higher prediction accuracy compared to Random Forest. Key factors influencing GDP were identified through feature correlation analysis. In conclusion, the combination of machine learning and metaheuristic-based optimization methods enhances GDP prediction accuracy, providing valuable insights for economic policymakers.