<![CDATA[The palm oil industry plays a strategic role in Indonesia's economy, yet challenges such as price fluctuations, market demand variability, and external factors like climate change often hinder accurate sales planning. This study aims to analyze palm oil sales patterns and compare the performance of the Random Forest and XGBoost algorithms in predicting sales at PT Geka Adhi Raksa. The methodology involves data preprocessing, model training, and prediction evaluation based on historical sales data from 2023 to 2024. Model performance is assessed using evaluation metrics such as RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), and R² (coefficient of determination). The results indicate that both algorithms provide reliable prediction accuracy, with XGBoost outperforming Random Forest in processing speed and overall accuracy. By adopting this data mining approach, the company can optimize sales strategies, avoid overstock or understock risks, and improve supply chain efficiency. This research is expected to serve as a reference for data-driven decision-making in agribusiness, particularly in the palm oil sector.]]>
Copyrights © 2025
