Model Regresi Mixed Geographically Weighted Compound Correlated Bivariate Poisson (Studi Kasus : Jumlah Kematian Ibu dan Jumlah Kematian Neonatal di Provinsi Jawa Timur Tahun 2022)

Pramesti, Era Ardhya (2024) Model Regresi Mixed Geographically Weighted Compound Correlated Bivariate Poisson (Studi Kasus : Jumlah Kematian Ibu dan Jumlah Kematian Neonatal di Provinsi Jawa Timur Tahun 2022). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003221015-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy

Abstract

Model Mix Geographically Weighted Compound Correlated Bivariate Poisson Regression (MGWCCBPR) merupakan kombinasi model Compound Correlated Bivariate Poisson Regression (CCBPR) dan model Geographically Weighted Compound Correlated Bivariate Poisson Regression (GWCCBPR). Model tersebut diusulkan untuk memberikan solusi yang tepat terhadap pemodelan data cacahan yang berkorelasi, dengan over-dispersi, kemencengan yang tinggi, dan modus yang lebih besar dari nol yang dipengaruhi oleh variabel global yang secara signifikan mempengaruhi semua pengamatan lokasi dan variabel lokal yang secara signifikan mempengaruhi pengamatan lokasi tertentu. Metode untuk memperoleh estimasi parameter pada model MGWCCBPR menggunakan Maximum Likelihood Estimation (MLE) dengan iterasi algoritma Bern, Hall, Hall, dan Hausman (BHHH). Sementara itu, pengujian hipotesis menggunakan metode Maximum Likelihood Ratio Test (MLRT). Selanjutnya, model MGWCCBPR diaplikasikan untuk memodelkan jumlah kasus kematian ibu dan neonatal di Provinsi Jawa Timur pada tahun 2022 dengan menggunakan enam variabel prediktor dan dua variabel eksposure, serta pembobot khusus yaitu kernel adaptive Gaussian. Dari penelitian ini diperoleh 38 model MGWCCBPR. Model MGWCCBPR untuk jumlah kematian ibu membagi kabupaten/kota di Provinsi Jawa Timur menjadi 3 kelompok dan model MGWCCBPR untuk jumlah kematian neonatal membagi kabupaten/kota di Provinsi Jawa Timur menjadi 2 kelompok, Variabel persentase kunjungan neonatal lengkap dan rata-rata lama sekolah perempuan merupakan variabel yang berpengaruh secara global di seluruh model MGWCCBPR baik untuk jumlah kematian ibu dan neonatal serta terdapat beberapa variabel lain yang hanya berpengaruh pada beberapa model MGWCCBPR untuk jumlah kematian ibu dan neonatal, dimana variabel-variabel tersebut dinamakan variabel yang berpengaruh secara lokal. Hasil penelitian ini juga menunjukkan bahwa model MGWCCBPR memiliki parsimoni model yang lebih baik jika dibandingkan dengan model GWCCBPR.
======================================================================================================================================
The Mix Geographically Weighted Compound Correlated Bivariate Poisson Regression (MGWCCBPR) model is a combination of the Compound Correlated Bivariate Poisson Regression (CCBPR) model and the Geographically Weighted Compound Correlated Bivariate Poisson Regression (GWCCBPR) model. The models are proposed to provide an appropriate solution to modeling correlated count data, with over-dispersion, high skewness, and mode greater than zero influenced by global variables that significantly affect all location observations and local variables that significantly affect location-specific observations. The method for obtaining parameter estimates in the MGWCCBPR model uses Maximum Likelihood Estimation (MLE) with the Bern, Hall, Hall, and Hausman (BHHH) iteration algorithm. Meanwhile, hypothesis testing uses the Maximum Likelihood Ratio Test (MLRT) method. Furthermore, the MGWCCBPR model was applied to model the number of maternal and neonatal death cases in East Java Province in 2022 using six predictor variables and two exposure variables, as well as special weights, namely the adaptive Gaussian kernel. From this study obtained 38 MGWCCBPR models for the number of maternal deaths and the number of neonatal deaths. The MGWCCBPR model for the number of maternal deaths divides districts / cities in East Java Province into 3 groups and the MGWCCBPR model for the number of neonatal deaths divides districts / cities in East Java Province into 2 groups, the variable percentage of complete neonatal visits and the average length of maternal schooling are variables that have a global effect in all MGWCCBPR models for both the number of maternal and neonatal deaths and there are several other variables that only affect some MGWCCBPR models for the number of maternal and neonatal deaths, where these variables are called locally influential variables. The results also show that the MGWCCBPR model has better parsimony model then GWCCBPR model.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	BHHH, Eksposure, MGWCCBPR, MLE, MLRT, Over-dispersi.
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA30.6 Spatial analysis H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HA Statistics > HA31.7 Estimation
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Era Ardhya Pramesti
Date Deposited:	29 Jul 2024 01:39
Last Modified:	29 Jul 2024 01:39
URI:	http://repository.its.ac.id/id/eprint/109179

Actions (login required)

View Item