A COMPARISON OF LOGISTIC REGRESSION, MIXED LOGISTIC REGRESSION, AND GEOGRAPHICALLY WEIGHTED LOGISTIC REGRESSION ON PUBLIC HEALTH DEVELOPMENT IN JAVA

Setiawan, Erwan
Suprayogi, Muhammad Azis
Kurnia, Anang

BAREKENG: Jurnal Ilmu Matematika dan Terapan
Vol 19 No 1 (2025): BAREKENG: Journal of Mathematics and Its Application

The Public Health Development Index (Indeks Pembangunan Kesehatan Masyarakat - IPKM) is a combined parameter that reflects progress in health development and is useful for determining areas that need assistance in improving health development. Through IPKM modeling, factors that significantly influence regional public health development can be discovered. This research aims to find an appropriate model for modeling IPKM and determine the factors that significantly influence public health development. The data used is the 2018 IPKM data collected from 119 cities/regencies in Java. We propose three models namely logistic regression (LR), mixed logistic regression (MLR), and geographically weighted logistic regression (GWLR). The research results show that the MLR is the best model for modeling IPKM in Java based on the AIC value criteria. Based on the MLR model, the factors that have a significant influence on public health development are the egg and milk consumption level and the percentage of the number of doctors per thousand population.