Under-five Mortality Rate Modeling Using Geographically and Temporal Weighted Regression (GTWR)

Abstract

Indonesia has a relatively high Under Five Mortality Rate. The causes of under-five mortality are low birth weight, pneumonia, education, socio-economic and cultural aspects. Central Java province is included in the province with high mortality rate in 2015 amounted to 11.64 per 1,000 live births (BPS, 2015). One of method to determine the factors who affected the under-five mortality using global regression analysis. Regression analysis of spatial data can be used as it has the nature of dependence between the measurement data by location. The local form of Regression analysis spatial is Geographically Weighted Regression (GWR). Development of methods GWR is Geographically and Temporal Weighted Regression method (GTWR). GTWR was accommodating spatial heterogeneity (location) and temporally (time). GTWR modeling using Gaussian Kernel functional Under Five mortality rates in central java province in 2013 - 2015 shows GTWR model is better than global regression method. Global regression has R² value 0.5080 and Mean Square Error (MSE) 27.137, while GTWR method has R² value of 0.5210 and MSE 1.084.