Forecasting Rice Production in Indonesia using Regression Techniques: A Comparative Analysis of Support Vector Machine, Linear Regression, and XGBoost Regression

Population growth is straining the agricultural sector, posing a significant threat to global food security. In Indonesia, rice is a staple diet, with annual demand growth of 1.16 percent outpacing production growth of 0.81 percent. Given Indonesia's projected population growth, meeting this demand is crucial. Indonesia is the fourth most populous nation in the world. In order to evaluate government policies on food sustainability, such as rice imports and agricultural programs, it is necessary to forecast rice production levels. Using Support Vector Machine, Linear Regression, and XGBoost Regression, the objective of this study was to forecast the paddy production in Indonesia. The study employed a dataset spanning the years 2018 to 2022, and the efficacy of each model was evaluated using standard metrics including MAPE, MAE, RMSE, and R2. With an RMSE of 0.015, an R2 of 0.997, and an MAE of 0.008, XGBoost Regression outperformed all other models across all metrics. Support Vector Regression had the highest MAPE, the lowest RMSE, the highest R2, and the lowest MAE, whereas Linear Regression had the highest MAPE, the lowest RMSE, the highest R2, and the lowest MAE.