MODEL REGRESI POISON BIVARIAT DENGAN KOVARIAN KONSTAN

DOI: https://doi.org/10.14710/medstat.11.1.27-38

Article Metrics: (Click on the Metric tab below to see the detail)

Bivariate Poisson models are appropriate for modeling paired count data exhibiting correlation. This study aims to estimates the parameters and test hypothesis of bivariate Poisson regression on modeling the number of infant mortality and maternal mortality in Central Java 2015. The parameters of the bivariate regression model are estimated by using the maximum likelihood method. Results show that the percentage of births by health personnel, the percentage of pregnant women administered the K4 program, the percentage of pregnant women receiving Fe3 tablets, percentage of exclusively breastfed infants, and percentage of households behaved in a clean and healthy life are significant for the number of infant mortality in Central Java. The variables that have significant effect on maternal mortality are percentage of births by health personnel, percentage of maternal women receiving postpartum health services, and percentage of pregnant women receiving Fe3 tablets. 

Keywords: Bivariate Poisson Regression, Infant Mortality, Maternal Mortality, Maximum Likelihood Estimation

  1. Untung Kurniawan 
    Badan Pusat Statistik Kabupaten Klaten, Indonesia