Model Regresi Bivariate Poisson Generalized Inverse Gaussian (Studi Kasus: Jumlah Kematian Ibu dan Neonatal di Jawa Tahun 2021)

Auliarahmi, Annisa (2024) Model Regresi Bivariate Poisson Generalized Inverse Gaussian (Studi Kasus: Jumlah Kematian Ibu dan Neonatal di Jawa Tahun 2021). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003221004-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 July 2026. Download (2MB) | Request a copy

Abstract

Regresi Poisson merupakan salah satu metode yang cocok digunakan untuk memodelkan data cacahan. Metode ini memiliki asumsi ketat yang harus dipenuhi yaitu rataan dan variansi bernilai sama (ekuidispersi) dimana pada kasus riil asumsi tersebut sulit untuk dipenuhi. Bivariate Poisson Generalized Inverse Gaussian Regression (BPGIGR) adalah salah satu pengembangan model regresi yang menggabungkan distribusi Poisson dengan distribusi Generalized Inverse Gaussian (GIG) untuk mengatasi kasus overdispersi dengan melibatkan dua variabel respon yang berkorelasi. GIG adalah distribusi kontinu dengan tiga parameter. Penelitian ini bertujuan untuk mendapatkan penaksir parameter dan statistik uji untuk model BPGIGR dengan menambahkan variabel eksposur. Variabel eksposur digunakan untuk mengatasi perbedaan ukuran populasi pada unit analisis dimana pada penelitian ini menyertakan dua variabel eksposur. Parameter model ditaksir menggunakan metode Maximum Likelihood Estimation (MLE) dengan metode numerik Berndt-Hall-Hall-Hausman (BHHH) yang tidak melibatkan turunan kedua. Pendekatan Maximum Likelihood Ratio Test (MLRT) digunakan untuk menurunkan statistik uji
=====================================================================================================================================
Poisson regression is a method that is suitable for modeling count data. This method has strict assumptions that must be met, namely the mean and variance have the same value (equidispersion), where in real cases this assumption is difficult to fulfill. Bivariate Poisson Generalized Inverse Gaussian Regression (BPGIGR) is a development of a regression model that combines the Poisson distribution with the Generalized Inverse Gaussian (GIG) distribution to overcome cases of overdispersion involving two correlated response variables. GIG is a continuous distribution with three parameters. This research aims to obtain parameter estimators and test statistics for the BPGIGR model by adding exposure variables. Exposure variables are used to overcome differences in population size in the analysis unit, which in this study includes two exposure variables. Model parameters are estimated using the Maximum Likelihood Estimation (MLE) method with the Berndt-Hall-Hall-Hausman (BHHH) numerical method which does not involve the second derivative. The Maximum Likelihood Ratio Test (MLRT) approach is used to derive test statistics

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Berndt-Hall-Hall-Hausman, Bivariate Poisson Generalized Inverse Gaussian Regression, maternal mortality, neonatal mortality.
Subjects:	H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HA Statistics > HA31.38 Data envelopment analysis. H Social Sciences > HA Statistics > HA31.7 Estimation H Social Sciences > HQ The family. Marriage. Woman R Medicine > RA Public aspects of medicine > RA971 Health services administration.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Annisa Auliarahmi
Date Deposited:	06 Feb 2024 07:11
Last Modified:	06 Feb 2024 07:11
URI:	http://repository.its.ac.id/id/eprint/106304

Actions (login required)

View Item