Model Mixed Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (Studi Kasus: Jumlah Kematian Ibu dan Neonatal di Provinsi Jawa Timur 2018-2022)

Terim, Mohammad Zanuar Fatih (2024) Model Mixed Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (Studi Kasus: Jumlah Kematian Ibu dan Neonatal di Provinsi Jawa Timur 2018-2022). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003222019-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (7MB) | Request a copy

Abstract

Distribusi Bivariate Poisson Inverse Gaussian (BPIG) merupakan distribusi mixed Poisson untuk dua variabel random berupa data cacahan yang saling berkorelasi serta mengalami overdispersi. Kajian distribusi BPIG pada pemodelan regresi spasial temporal menghasilkan model Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (GTWBPIGR). Dalam pemodelan spasial temporal, tidak semua variabel prediktor berpengaruh secara lokal, seringkali terdapat variabel prediktor yang berpengaruh secara global. Maka, dalam penelitian ini akan dilakukan pengembangan dari model GTWBPIGR menjadi Model Mixed Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (MGTWBPIGR). Model MGTWBPIGR adalah model regresi spasial temporal yang mempertimbangkan adanya variabel lokal dan global sehingga akan menghasilkan variabel prediktor yang berpengaruh secara lokal dan global pada setiap lokasi dan waktu pengamatan. Variabel eksposur akan dilibatkan dalam pemodelan MGTWBPIGR sehingga setiap pengamatan layak untuk dibandingkan satu sama lain. Estimasi parameter model MGTWBPIGR dilakukan dengan menggunakan Maximum Likelihood (MLE) dengan algoritma Berndt-Hall-Hall-Hausman (BHHH). Pengujian hipotesis menggunakan Maximum Likelihood Ratio Test (MLRT). Model MGTWBPIGR akan diterapkan pada data jumlah kematian ibu dan jumlah kematian neonatal di Provinsi Jawa Timur tahun 2018-2022. Berdasarkan hasil penelitian menggunakan model MGTWBPIGR, variabel prediktor yang bersifat global adalah persentase kunjungan ibu hamil K4, persentase ibu hamil mendapat imunisasi TD+, dan persentase penanganan komplikasi neonatal sedangkan variabel predictor lainnya berpengaruh secara lokal. Selanjutnya pada model MGTWBPIGR, jumlah kematian ibu dan jumlah kematian neonatal terbagi menjadi dua kelompok berdasarkan variabel prediktor yang signifikan pada setiap kabupaten/kota.
====================================================================================================================================
Bivariate Poisson Inverse Gaussian (BPIG) distribution is a mixed Poisson distribution for two random variables in the form of chopped data that are correlated and overdispersed. The study of the BPIG distribution in spatial temporal regression modeling resulted in the Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (GTWBPIGR) model. In spatial temporal modeling, not all predictor variables affect locally, there are often predictor variables that affect globally. Therefore, this study will develop the GTWBPIGR model into the Mixed Geographically and Temporally Weighted Bivariate Poisson Inverse Gaussian Regression (MGTWBPIGR) model. The MGTWBPIGR model is a spatial temporal regression model that considers local and global variables so that it will produce predictor variables that have local and global effects at each location and time of observation. Exposure variables will be involved in MGTWBPIGR modeling so that each observation is feasible to compare with each other. Parameter estimation of the MGTWBPIGR model is done using Maximum Likelihood (MLE) with the Berndt-Hall-Hausman (BHHH) algorithm. Hypothesis testing using the Maximum Likelihood Ratio Test (MLRT). The MGTWBPIGR model will be applied to data on the number of maternal deaths and the number of neonatal deaths in East Java Province in 2018-2022. Based on the results of research using the MGTWBPIGR model, the predictor variables that are global are the percentage of visits to K4 pregnant women, the percentage of pregnant women receiving TD+ immunization, and the percentage neonatal complication handled while other predictor variables have a local effect. Furthermore, in the MGTWBPIGR model, the number of maternal deaths and the number of neonatal deaths are divided into two groups based on significant predictor variables in each district/city.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	BHHH, BPIGR, MGTWBPIGR, Jumlah Kematian Ibu, Jumlah Kematian Neonatal. ============================================================ BHHH, BPIGR, MGTWBPIGR, Number of Maternal Deaths, Number of Neonatal Deaths.
Subjects:	H Social Sciences > HA Statistics > HA30.6 Spatial analysis Q Science > QA Mathematics > QA371 Differential equations--Numerical solutions
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Mohammad Zanuar Fatih Terim
Date Deposited:	24 Jan 2025 06:30
Last Modified:	24 Jan 2025 06:30
URI:	http://repository.its.ac.id/id/eprint/116807

Actions (login required)

View Item