Pemodelan Regresi Hurdle Negative Binomial Tersensor Pada Kasus Penyakit Difteri

Abstract

The censored Hurdle Negative Binomial regression model is a method which can be used for response variables of count data with excess zero values. This model is an alternative solution to overcome over dispersion in Poisson regression due to many zero-value observations. This model uses a two-part approach, the first part is modeling the response variable which a zero value and the second part is modeling the response variable with a non-negative integer value. The type of censored used in this research is the right censored. This research will applied on the theory and empirical studies on the censored Hurdle Negative Binomial regression model. the results of the theory estimation of the parameters of the censored Hurdle Negative Binomial regression model using the maximum likelihood method are obtained implicit nonlinear equations, so the estimation of the model parameters using Newton Rapshon method. Censored Hurdle Negative Binomial regression model of diphtheria cases in Riau Province with the variables of basic immunization coverage, coverage of toddler health services and coverage of health facilities affect the number of diphtheria cases.