Factors Affecting the Number of Infant Morality Cases in West Java for the 2019-2020 Period using Generalized Poisson Regression (GPR)

Kartika Dewi (Department of Mathematics, Faculty of Mathematics and Natural Science (FMIPA), Universitas Padjadjaran, Bandung, Indonesia)
Nurul Gusriani (Department of Mathematics, Faculty of Mathematics and Natural Science (FMIPA), Universitas Padjadjaran, Bandung, Indonesia)
Kankan Parmikanti (Department of Mathematics, Faculty of Mathematics and Natural Science (FMIPA), Universitas Padjadjaran, Bandung, Indonesia)

Eksakta : Berkala Ilmiah Bidang MIPA
Vol. 24 No. 02 (2023): Eksakta : Berkala Ilmiah Bidang MIPA (E-ISSN : 2549-7464)

The number of infant mortality cases is data in the form of counts which is modeled by Poisson regression. There is an assumption that needs to be met, namely equidispersion. Equidispersion is a condition in which the mean and variance of the variables are the same, but in practice this assumption is often not met. There are two possible events, namely overdispersion and underdispersion. The Generalized Poisson Regression (GPR) model is one solution to solve this problem. In estimating the GPR parameter, the Maximum Likelihood Estimation (MLE) method is used, but the derivation of the log-likelihood function does not always produce explicit results, so the Newton-Raphson iteration method is used. Poisson regression analysis conducted on the number of infant mortality cases in West Java showed that the model had overdispersion as seen from the value of the dispersion parameter which was more than zero, so the GPR model was used. Parameter significance test was carried out on three factors, namely the poverty gap index , the percentage of low birth weight infants , and the percentage of exclusive breastfeeding for infants  the results obtained that all factors affected the number of infant mortality cases in West Java.