Comparison of Rainfall Prediction Results in South Bangka Regency Using Support Vector Regression and SARIMA

Rainfall is one of the key climate variables. It plays an important role in the hydrological cycle. In the field of agriculture, it is crucial as it determines the availability of water for crops, helping farmers manage water scarcity issues. South Bangka is one of the districts with the highest rice production levels compared to other districts in the Bangka Belitung Islands Province. This study aims to predict the rainfall in South Bangka for the next four years, from January 2024 to December 2027, using the Support Vector Regression (SVR) and SARIMA methods. The results of this study indicate that the SVR method is the best for prediction compared to the SARIMA method, with an average MAPE value of 0.03%. The kernel used is the Radial Basis Function (RBF), with parameter values including epsilon (ԑ) of 0.0001, Cost (C) of 1000, and gamma (γ) of 235. The MAPE value for the training data is 0.045%, and for the test data, it is 0.015%. The best SARIMA model is (3,1,3)(1,1,3)24 with a MAPE value of 15.51%.