Parameter estimation and statistical test on bivariate zero inflated Poisson Inverse Gaussian with exposure variable.

This research focused on parameter estimation and hypothesis testing of Bivariate Zero-Inflated Poisson Inverse Gaussian Regression (BZIPIGR) type II. BZIPIGR model is a mixed Poisson distribution used to model data with excess-zero and overdispersion. The parameter estimation in this study used the Maximum Likelihood Estimation (MLE) method. The first derivative of the log-likelihood function from the BZIPIGR model had obtained not closed-form. Therefore, to maximize the likelihood function, it had used numerical iteration, namely the Berndt-Hall-Hall-Hausman (BHHH). The determination of the critical area and the significance of the model parameters had used the Maximum Likelihood Ratio Test (MLRT). [ABSTRACT FROM AUTHOR]