Penaksiran Parameter Pada Model Generalized Poisson Spatial Autoregressive (Studi Kasus: Pemodelan Jumlah Kasus Pneumonia pada Balita di Kabupaten Tuban)

Sihombing, Joshua Capri Gunawan (2026) Penaksiran Parameter Pada Model Generalized Poisson Spatial Autoregressive (Studi Kasus: Pemodelan Jumlah Kasus Pneumonia pada Balita di Kabupaten Tuban). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003232015-Master_Thesis.pdf Download (3MB)

Abstract

Jumlah kasus pneumonia pada balita di Kabupaten Tuban menyajikan dua tantangan data yang signifikan, yaitu overdispersi dan dependensi spasial. Penelitian ini bertujuan untuk mengembangkan dan menerapkan model Generalized Poisson Spatial Autoregressive (GPSAR) untuk mengatasi kedua masalah tersebut secara bersamaan. Estimasi parameter model dilakukan menggunakan metode Maximum Likelihood Estimation dengan prosedur iterasi Berndt-Hall-Hall-Hausman (BHHH). Hasil penelitian mengonfirmasi validitas dan keunggulan model spasial yang diajukan. Model GPSAR terbukti lebih unggul dibandingkan model non-spasial Generalized Poisson Regression (GPR) dalam hal kebaikan model (AICc: 1673,39 vs 1717,24). Secara statistik, parameter struktural untuk lag spasial dan parameter dispersi bernilai signifikan yang mengonfirmasi adanya pengelompokan spasial yang kuat dan fenomena overdispersi pada data. Tiga variabel prediktor ditemukan berpengaruh signifikan: Cakupan Air Bersih (X4) yang menunjukkan efek protektif (negatif), serta Kelas Ibu Hamil (X3) dan Imunisasi Campak (X5) yang menunjukkan asosiasi positif yang berpotensi berkaitan dengan kualitas surveilans kesehatan.
======================================================================================================================================
The number of pneumonia cases in children under five in Tuban Regency presents two significant data challenges: overdispersion and spatial dependency. This study aims to develop and apply the Generalized Poisson Spatial Autoregressive (GPSAR) model to address both issues simultaneously. The model parameters were estimated using the Maximum Likelihood Estimation method with the Berndt-Hall-Hall-Hausman (BHHH) iteration procedure. The results confirm the validity and superiority of the spatial model. The GPSAR model outperformed the non-spatial Generalized Poisson Regression (GPR) model in terms of goodness-of-fit (AICc: 1673,39 vs 1717,24). Statistically, the structural parameters for spatial lag and dispersion were highly significant , confirming strong spatial clustering and overdispersion in the data. Three predictor variables were found to be statistically significant determinants: Clean Water Coverage (X4), which showed a protective effect, as well as Prenatal Classes (X3) and Measles Immunization (X5), which showed positive associations potentially linked to healthcare surveillance quality.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Data Cacah, Model SAR, Overdispersi, Pneumonia, Spatial Lag, Count Data, SAR Model, Overdispersion, Pneumonia, Spatial Lag.
Subjects:	H Social Sciences > HA Statistics > HA30.6 Spatial analysis
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Joshua Capri Gunawan Sihombing
Date Deposited:	31 Jan 2026 05:35
Last Modified:	31 Jan 2026 05:35
URI:	http://repository.its.ac.id/id/eprint/131385

Actions (login required)

View Item