<![CDATA[Poisson regression may be used to describe count data in the form of positive integers as it often follows a Poisson distribution. Poisson regression requires that the response variable's mean and variance be equal (equidispersion). In practice, however, it is more typical to find data with overdispersion, or variation larger than the mean. The number of newborn and maternal fatalities are the response variables, and the units of analysis are the East Java Province's regencies and cities. Because of the correlation between these two variables and the overdispersion that results, the Poisson regression model has to be further developed. One such model development that blends the Poisson and Lognormal distributions is called Bivariate Poisson Lognormal Regression (BPLNR). In order to predict the factors thought to be impacting the number of infant deaths and maternal deaths in East Java Province in 2021, this study attempts to produce parameter estimators and test data for the BPLNR model. According to the modeling results, the number of infant and maternal fatalities is significantly impacted by a variety of variables, including the proportion of cases treated by health personnel and the percentage of K4 antenatal visits by expectant women, among others. Furthermore, the dispersion parameter indicates that overdispersion in the data on newborn and maternal mortality in East Java in 2021 has been taken into account by the model.]]>
Copyrights © 2026
