Global climate change and rising extreme weather events necessitate mitigation efforts using precise data. This community service program introduces a rainfall prediction model based on the Tweedie Mixture with PCA reduction at BMKG Lampung. The model addresses asymmetric rainfall data, which often contains many zero values, while optimizing predictor data through PCA-based dimension reduction. Conducted at BMKG’s Climatology Station in Lampung, the program engaged employees and technical staff through workshops, training, and prediction simulations. Participants learned the Tweedie Mixture and PCA theory, statistical software usage, and model applications for climate risk mitigation, such as planting schedules and water management. Evaluations show the Tweedie Mixture outperforms conventional methods in accuracy, enhancing BMKG’s climate analysis and weather forecasting.
                        
                        
                        
                        
                            
                                Copyrights © 2025