<![CDATA[Weather uncertainty due to climate change increasingly threatens rice harvest success, especially when farmers still rely on traditional forecasts that are not always accurate. This study developed a decision support system to determine the timing of rice planting based on daily rainfall predictions using Random Forest Regression. Daily climate data from the BMKG in Tuban, East Java, for the period 2022–2025 was used as the basis for training, with the addition of time features such as month, day of the year, and season to capture seasonal patterns. In East Java, the rainy season usually lasts from October to April and the dry season from May to September, but climate change has caused shifts in the timing, duration, and intensity of rainfall, making traditional seasonal classifications less reliable for determining the optimal planting time. The model was tested on 2025 data and showed improved performance compared to the baseline model. The tuned model produced an MAE of 5.78 mm, an RMSE of 9.75 mm, and a coefficient of determination (R²) of 0.177, an improvement over the baseline, which had an MAE of 6.02 mm, an RMSE of 10.16 mm, and an R² of 0.107. Although the R² value is still relatively low, the decrease in MAE and RMSE indicates that the tuned model is more accurate in predicting daily rainfall, especially in the light to moderate range, which is most relevant for planting decisions]]>
Copyrights © 2026
