<![CDATA[Accurate rainfall prediction is crucial for supporting the agricultural sector in Lampung Province. This research employs the Exponential Dispersion Model (EDM), a special case of the Generalized Linear Model (GLM), incorporating a Tweedie mixture distribution with Principal Component Analysis (PCA) to reduce correlated variables. Rainfall data were obtained from the Meteorology, Climatology, and Geophysics Agency (BMKG) through twelve rain observation posts (2013-2022), and supplemented with precipitation data from the General Circulation Model (GCM) obtained from the European Centre for Medium-Range Weather Forecasts (ECMWF). The Tweedie mixture distribution was selected for its ability to handle non-normally distributed rainfall data containing zero values. The results show that the Root Mean Square Error of Prediction (RMSEP) for the Tweedie mixture-PCA model at the Gisting Atas station is 163.90, while the Normal-PCA model achieved 169.11. Therefore, the Tweedie mixture-PCA approach is more effective and recommended for improving rainfall prediction in Lampung Province, offering potential benefits for agricultural planning and resource management.]]>
Copyrights © 2025
