<![CDATA[Palm oil milling operations are frequently affected by fluctuating fresh fruit bunch (FFB) availability, variations in Oil Extraction Rate (OER), and limitations in production planning under dynamic operational and environmental conditions. In many mills, production decisions still rely heavily on operator experience and manually recorded information, resulting in inconsistent and less structured planning processes. This study develops an interpretable Artificial Intelligence-based Decision Support System (AI-DSS) that integrates a Random Forest Regression model into a web-based decision support platform to support enterprise-level palm oil production planning. The predictive model was developed using operational and environmental data collected from a palm oil mill and evaluated through 10-fold cross-validation. The proposed Random Forest Regression model achieved an average coefficient of determination (R²) of 0.94 ± 0.02 and an average Root Mean Square Error (RMSE) of 1.06 ± 0.09, demonstrating strong predictive capability for modeling nonlinear relationships among production variables. The trained Random Forest model was integrated into a web-based decision support framework that transforms predicted OER values into structured operational recommendations while preserving managerial decision-making. Feature importance analysis further enhances system interpretability by identifying Fresh Fruit Bunch (FFB) intake and operational hours as the most influential variables affecting OER. The proposed AI-DSS demonstrates the practical feasibility of integrating predictive analytics, explainable machine learning, and enterprise-level decision support to facilitate more structured, transparent, and data-driven production planning in palm oil mill operations.]]>
Copyrights © 2026
