Model Geographically Weighted Bivariate Poisson Generalized Inverse Gaussian Regression (Jumlah Kematian Ibu Nifas dan Post Neonatal di Jawa Timur Tahun 2023)

Leiwakabessy, Reyner Marvi (2025) Model Geographically Weighted Bivariate Poisson Generalized Inverse Gaussian Regression (Jumlah Kematian Ibu Nifas dan Post Neonatal di Jawa Timur Tahun 2023). Masters thesis, Institut Tegnologi Sepuluh Nopember.

Text 6003222003-Mater_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (3MB) | Request a copy

Abstract

Regresi Poisson adalah analisis regresi nonlinear yang digunakan untuk menganalisis data diskrit dan terdapat asumsi bahwa rata-rata dan variansi dari variabel respon harus sama (equidispersi). Namun, dalam praktiknya, sering terjadi pelanggaran asumsi ini yang dikenal sebagai underdispersi atau overdispersi. Bivariate Poisson Generalized Inverse Gaussian Regression (BPGIGR) adalah jenis regresi mixed Poisson yang dapat digunakan untuk memodelkan data count berpasangan yang mengalami overdispersi. Penelitian ini mengembangkan model Geographically Weighted Bivariate Poisson Generalized Inverse Gaussian (GWBPGIGR) untuk secara simultan menangani heterogenitas spasial. Estimasi parameter model GWBPGIGR dilakukan menggunakan Maximum Likelihood Estimation (MLE) dengan algoritma Berndt-Hall-Hall-Hausman (BHHH). Pengujian parameter secara serentak dilakukan menggunakan Maximum Likelihood Ratio Test (MLRT), sedangkan pengujian parsial dilakukan menggunakan uji Z. Hasil penelitian menunjukkan bahwa pada model BPGIGR, variabel prediktor yang berpengaruh signifikan terhadap kematian ibu nifas adalah persentase ibu hamil melaksanakan kunjungan pertama (K1), persentase ibu hamil melaksanakan kunjungan keempat (K4), dan persentase ibu hamil mendapat tablet tambah darah (TTD), sedangkan untuk jumlah kematian bayi post neonatal adalah persentase ibu hamil melaksanakan kunjungan keempat (K4) dan persentase persalinan oleh tenaga kesehatan. Selanjutnya pada model GWBPGIGR, jumlah kematian ibu nifas membagi kabupaten/kota menjadi 4 kelompok dan untuk jumlah kematian bayi post neonatal membagi kelompok kabupaten/kota menjadi 2 kelompok. Berdasarkan nilai AICc, model GWBPGIGR lebih baik dari model BPGIGR dalam memodelkan data jumlah kematian ibu nifas dan kematian bayi post neonatal di Provinsi Jawa Timur pada Tahun 2023.
==============================================================================================================================
Poisson regression is a nonlinear regression analysis used to analyze discrete data and there is an assumption that the mean and variance of the response variable must be equal (equidispersion). However, in practice, there is often a violation of this assumption known as underdispersion or overdispersion. Bivariate Poisson Generalized Inverse Gaussian Regression (BPGIGR) is a type of mixed Poisson regression that can be used to model paired count data that experience overdispersion. This study develops the Geographically Weighted Bivariate Poisson Generalized Inverse Gaussian (GWBPGIGR) model to simultaneously handle spatial heterogeneity. Parameter estimation of the GWBPGIGR model was conducted using Maximum Likelihood Estimation (MLE) with the Berndt-Hall-Hausman (BHHH) algorithm. Simultaneous parameter testing was conducted using the Maximum Likelihood Ratio Test (MLRT), while partial testing was conducted using the Z test. The results showed that in the BPGIGR model, the predictor variables that had a significant effect on postpartum mortality were the percentage of pregnant women conducting the first visit (K1), the percentage of pregnant women conducting the fourth visit (K4), and the percentage of pregnant women receiving blood supplement tablets (TTD), while for the number of post neonatal infant deaths were the percentage of pregnant women conducting the fourth visit (K4) and percentage of deliveries by health workers. Furthermore, in the GWBPGIGR model, the number of postpartum deaths divides districts/cities into 4 groups and for the number of post neonatal infant deaths divides districts/cities into 2 groups. Based on the AICc value, the GWBPGIGR model is better than the BPGIGR model in modeling data on the number of maternal deaths and post neonatal infant deaths in East Java Province in 2023.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	BHHH, GWBPGIGR, MLE, Kematian ibu nifas, Maternal mortality, post neonatal
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation. Logistic regression analysis. H Social Sciences > HA Statistics > HA31.7 Estimation Q Science > Q Science (General) Q Science > Q Science (General) > Q180.55.M38 Mathematical models
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Leiwakabessy Reyner Marvi
Date Deposited:	03 Feb 2025 13:46
Last Modified:	03 Feb 2025 13:46
URI:	http://repository.its.ac.id/id/eprint/117957

Actions (login required)

View Item