COMPARISON OF GLM, GLMM AND GEE POISSON MATHEMATICAL MODELING PERFORMANCE (Case Study: Number of Pulmonary Tuberculosis Patients in Indonesia in 2019-2021)

Abstract

This study aims to compare the performance of data modeling with Poisson regression with Generalized Linear Model (GLM), Generalized Linear Mixed Model (GLMM), and Generalized Estimating Equation (GEE) modeling. The case study used is a factor that affects the number of Pulmonary Tuberculosis cases in Indonesia with panel data. Based on the AIC criteria, the smallest BIC and RMSE GLMM models perform better than GLM and GEE. In addition, GLMM modeling also has a coefficient of determination value. The results showed that the percentage of the population smoking and the percentage of Unmet Kesehatan had a significant positive influence on the number of Pulmonary Tuberculosis cases. In contrast, the percentage of households who had access to handwashing with soap and proper drinking water significantly negatively influenced the number of Pulmonary Tuberculosis cases. Therefore, the public needs awareness to reduce cigarette consumption and increase environmental access, such as adequate and clean drinking water sources and handwashing.