Geographically Weighted Regression with Bi-Square Kernel Weights for Life Expectancy Data in East Java Province

Abstract

This study explores the application of Geographically Weighted Regression (GWR) using a Bi-Square Kernel weighting function to analyze life expectancy data across East Java Province. By incorporating spatial heterogeneity, the GWR model provides more accurate and localized insights compared to traditional global regression models. The results indicate significant spatial variability in the effects of poverty rate, healthcare facilities, sanitation, health complaints, and immunization coverage on life expectancy. Based on the analysis of life expectancy estimates in regencies/cities of East Java Province, the Madura region exhibits lower life expectancy compared to other areas, with Bangkalan Regency having the lowest life expectancy at 61.43 years. Additionally, urban areas generally have higher life expectancy than rural areas, with Surabaya City recording the highest life expectancy in East Java at 72.03 years. This disparity can be attributed to differences in the quality of healthcare services and better access to healthcare in urban areas compared to rural ones.