Safarida, Rizki (2023) Model Regresi Geographically Weighted Compound Correlated Bivariate Poisson (Studi Kasus: Jumlah Kematian Ibu Nifas dan Jumlah Kematian Neonatal di Provinsi Jawa Timur Tahun 2021). Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6003211033-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6003211033-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy |
Abstract
Pada penelitian yang berbasis wilayah dapat dimungkinkan terjadi adanya efek spasial yaitu berupa heterogenitas spasial. Adanya heterogenitas spasial menyebabkan pemodelan yang bersifat global menjadi kurang tepat, sehingga perlu dilakukan pengembangan model. Compound Correlated Bivariate Poisson Regression (CCBPR) adalah salah satu pengembangan model regresi bivariate Poisson untuk data respon berupa cacahan yang mengalami overdispersi dan memiliki skewness tinggi dengan modus lebih besar nol. Model tersebut merupakan model global sehingga dilakukan pengembangan model yaitu Geographically Weighted CCBPR (GWCCBPR) yang dapat mengakomodir adanya heterogenitas spasial. Pada penelitian ini, model GWCCBPR diaplikasikan pada kasus jumlah kematian ibu nifas dan jumlah kematian neonatal di Provinsi Jawa Timur tahun 2021. Penaksiran parameter model GWCCBPR didapatkan dengan metode Maximum Likelihood Estimation (MLE) dengan iterasi Berndt Hall Hall Hausman (BHHH). Pengujian hipotesis menggunakan metode Maximum Likelihood Ratio Test (MLRT). Model GWCCBPR menghasilkan nilai AICc yang lebih rendah daripada model CCBPR. Hal tersebut berarti bahwa model GWCCBPR lebih baik dalam memodelkan jumlah kasus kematian ibu nifas dan jumlah kasus kematian neonatal di Provinsi Jawa Timur tahun 2021. Pemodelan GWCCBPR menghasilkan dua (2) kelompok kabupaten/kota berdasarkan variabel prediktor yang signifikan terhadap variabel respon
=================================================================================================================================
In research based on geographical regions, spatial effects such as spatial heterogeneity can occur. The existence of spatial heterogeneity causes global modeling to be less accurate, necessitating model development. Compound Correlated Bivariate Poisson Regression (CCBPR) is one of the model developments of bivariate Poisson regression for count data with overdispersion and high skewness, with mode greater than zero. This model is a global model, so model development called Geographically Weighted CCBPR (GWCCBPR) is performed to accommodate spatial heterogeneity. In this study, the GWCCBPR model is applied to the cases of postnatal maternal mortality and neonatal mortality in East Java Province in 2021. The parameter estimation of the GWCCBPR model is obtained using the Maximum Likelihood Estimation (MLE) method with Berndt Hall Hall Hausman (BHHH) iteration. Hypothesis testing is conducted using the Maximum Likelihood Ratio Test (MLRT) method. The GWCCBPR model yields a lower AICc value than the CCBPR model. This means that the GWCCBPR model is better at modeling the number of postnatal maternal mortality and number of neonatal mortality cases in East Java Province in 2021. The GWCCBPR modeling results in two (2) groups of districts/cities based on significant predictor variables related to the response variable
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | GWCCBPR, Jumlah Kematian Ibu Nifas, Jumlah Kematian Neonatal, Number of Postnatal Maternal Mortality, Number of Neonatal Mortality |
| Subjects: | H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA30.6 Spatial analysis |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Rizki Safarida |
| Date Deposited: | 02 Oct 2023 03:46 |
| Last Modified: | 02 Oct 2023 03:46 |
| URI: | http://repository.its.ac.id/id/eprint/104423 |
Actions (login required)
 |
View Item |