Machine Learning-Based Prediction of Palm Oil Production in Merangin Regency

Palm oil production in Merangin Regency experiences fluctuations influenced by various factors, making a prediction model necessary to support planning and decision-making. This study aims to develop a prediction model for palm oil production using Multiple Linear Regression and Random Forest methods. The data used are secondary data obtained from the Plantation Office of Merangin Regency for the period 2016–2025, which include productive plantation area, plant age, number of farmers, sub-districts, and previous production.The Multiple Linear Regression method is used to analyze the relationships between variables, while Random Forest is used for prediction. Model evaluation is carried out using Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) metrics with an 80% training and 20% testing data split.The results show that Random Forest outperforms Multiple Linear Regression, with an MAE value of 0.60, MSE of 324.07, and RMSE of 18.00. The previous production variable is the most dominant factor influencing production outcomes. The developed model has also been successfully implemented in a simple application-based prediction system.Thus, the Random Forest method is proven to be effective in predicting palm oil production and can be used as a decision-support tool in the plantation sector.