Rainfall is one of the crucial meteorological elements that can significantly impact human life. Accurate rainfall prediction is essential for effective natural resource planning and management across various regions, especially in Java Island, which is one of the most densely populated areas in Indonesia. However, predicting rainfall distribution in this region is challenging due to its complex climate patterns. This study aims to address this challenge by developing a rainfall distribution prediction model for Java Island using Support Vector Machine (SVM). The model incorporates time-based feature expansion to enhance the predictive capability of the SVM. Additionally, the method is combined with Kriging interpolation to accurately classify the spatial distribution of rainfall across Java Island. The model was trained and tested using monthly data from 27 meteorological stations, covering the period from January 1, 2020, to March 31, 2022. The results demonstrate that the model's performance exceeds 90%, indicating its effectiveness in predicting future rainfall distribution classifications. The contribution of this research lies in providing insights into feature expansion techniques in machine learning, refining predictive models applied in meteorology and environmental management, and addressing the complexity of rainfall prediction in densely populated regions.
Keywords: Support Vector Machine; time-based feature expansions; rainfall; classification prediction; interpolation krigging;