Parameter Estimation and Statistical Test on Zero Inflated Poisson Inverse Gaussian Regression Model

Ermawati*, Purhadi, Santi Puteri Rahayu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The over dispersion case is one of the assumptions that violated in the data count model using Poisson regression, which was caused by the presence of many zero values (more than 30%) on the response variable. Zero-Inflated Poisson Inverse Gaussian Regression (ZIPIGR) is a method that can use to model this case. In this research, the parameter estimation of the ZIPIGR model used the Maximum Likelihood Estimation (MLE) method. But from the derivatives process, it was found that the first derivatives of log likelihood from the ZIPIGR model are not close form. Therefore Berndt-Hall-Hall-Hallman (BHHH) iteration is used to facilitate MLE for obtained parameter estimation. Statistical test of the ZIPIGR model obtained using the Maximum Estimated Ratio Test (MLRT).

Original languageEnglish
Title of host publication8th International Conference and Workshop on Basic and Applied Science, ICOWOBAS 2021
EditorsAnjar Tri Wibowo, M. Fariz Fadillah Mardianto, Riries Rulaningtyas, Satya Candra Wibawa Sakti, Muhammad Fauzul Imron, Rico Ramadhan
PublisherAmerican Institute of Physics Inc.
ISBN (Electronic)9780735442610
DOIs
Publication statusPublished - 25 Jan 2022
Event8th International Conference and Workshop on Basic and Applied Science, ICOWOBAS 2021 - Surabaya, Indonesia
Duration: 25 Aug 202126 Aug 2021

Publication series

NameAIP Conference Proceedings
Volume2554
ISSN (Print)0094-243X
ISSN (Electronic)1551-7616

Conference

Conference8th International Conference and Workshop on Basic and Applied Science, ICOWOBAS 2021
Country/TerritoryIndonesia
CitySurabaya
Period25/08/2126/08/21

Fingerprint

Dive into the research topics of 'Parameter Estimation and Statistical Test on Zero Inflated Poisson Inverse Gaussian Regression Model'. Together they form a unique fingerprint.

Cite this