ZERO INFLATED POISSON REGRESSION MODELS TO ANALYZE FACTORS THAT INFLUENCE THE NUMBER OF MEASLES CASE IN JAVA

Abstract

Measles is an infectious disease that often occurs in children and is caused by the measles virus (morbillivirus) which can cause death. Thus, it is important to identify the factors that cause measles. The number of measles cases is used as response variable in the form discrete data so that Poisson Regression is commonly used. However, some assumptions are sometimes not met, such as overdispersion and excess zero so that can use Zero Inflated Poisson Regression to meet these assumptions. Because the model can overcome two common characteristics that are often found in count data, which are excess zero and overdispersion. The purpose of this study was to determine the factors that influence the number of measles cases in East Java. The data in the study used secondary data obtained from the Central Statistics Agency (BPS). The predictor variables used were the number of population, percentage of vaccination, percentage of poor people, and percentage of adequate sanitation. The results showed that the data is overdispersed because the variance is greater than the mean. There were four predictor variables, The -value of the total population variable is <0.01, the percentage of vaccinations is 0.914, the percentage of poor people <0.01 and the percentage of proper sanitation is 0.014 so it can be concluded that the percentage of vaccinations has no effect on the number of measles cases and the other three variables affect the number of measles cases in East Java. The best model of affect the number of measles cases in East Java is Zero Inflated Poisson with AIC value 326.24. The ZIP model for measles case in East Java is .

Keywords

• excess zero
• measles
• overdispersion
• poisson regression
• zero inflated poisson regression