Integrating generalized linear mixed models with extreme neural network: enhancing pulmonary tuberculosis risk modeling in West Java, Indonesia

Abstract

Mycobacterium tuberculosis is the primary cause of tuberculosis (TB), a global public health concern that primarily affects the respiratory system. This disease is also prevalent in the West Java Province of Indonesia. The study aims to analyze tuberculosis (TB) patterns using extreme values to identify the most accurate models for forecasting illness cases. The study uses advanced machine learning methods, the Generalized Linear Mixed Model (GLMM), and the Extreme Neural Network (ENN), to investigate how environmental and personal factors affect the number of tuberculosis (TB) cases. Results indicate that the population density and age groups of 45–64 and over 65 years strongly influence the occurrence of tuberculosis in West Java. We utilize the FFNN model to predict the future number of TB cases and risk variables. To effectively prevent and manage the spread of tuberculosis in the community, it is crucial for all parties to be watchful and aware of the different risk factors linked to the disease, as revealed by the findings.