Nabila, Galuh Putri (2025) Perbandingan Metode Geographically Weighted Generalized Poisson Regression Dan Geographically Weighted Negative Binomial Regression Dalam Analisa Faktor Kasus Tuberculosis (TBC). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002211071-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002211071-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Tuberculosis (TBC) merupakan salah satu penyakit menular yang menjadi permasalahan kesehatan global, termasuk di Indonesia. Jawa Timur memiliki jumlah kasus TBC yang cukup tinggi, sehingga diperlukan analisis mendalam mengenai faktor-faktor yang memengaruhi penyebarannya. Beberapa penelitian sebelumnya telah menerapkan model spasial seperti Geographically Weighted Poisson Regression (GWPR) dan Geographically Weighted Regression (GWR), namun masih terbatas pada analisis hubungan global tanpa mempertimbangkan distribusi spasial yang lebih kompleks.
Penelitian ini bertujuan untuk membandingkan metode Geographically Weighted Generalized Poisson Regression (GWGPR) dan Geographically Weighted Negative Binomial Regression (GWNBR) dalam menganalisis faktor-faktor yang memengaruhi penyebaran kasus TBC dengan mempertimbangkan aspek spasial dan overdispersion. Hasil penelitian menunjukkan bahwa pada tahun 2023 terdapat 85.060 kasus TBC di Provinsi Jawa Timur dengan rata-rata 2238,421 kasus per Kota/Kabupaten dan standar deviasi sebesar 1885,143, yang mengindikasikan persebaran data yang cukup besar. Pemodelan dengan metode GWGPR menghasilkan empat kelompok wilayah berdasarkan variabel signifikan, sedangkan GWNBR menghasilkan lima kelompok. Kedua model, GWGPR dan GWNBR, menghasilkan estimasi parameter yang bervariasi untuk setiap
Kota/Kabupaten, mencerminkan kondisi nyata di lapangan. Berdasarkan nilai Mean Square Error (MSE) dan Corrected Akaike Information Criterion (AICc), model GWGPR dinilai lebih baik dengan nilai MSE sebesar 7976165,036 dan AICc sebesar 613,1458. Uji Breusch-Pagan menunjukkan adanya heterogenitas spasial, sedangkan uji Moran’s I menunjukkan tidak adanya ketergantungan spasial, sehingga pembobot spasial kernel adaptive bisquare sesuai diterapkan pada kedua model. Dengan demikian, metode GWGPR paling sesuai untuk memodelkan jumlah kasus TBC di Kota/Kabupaten di
Jawa Timur dibandingkan metode GWNBR.
============================================================================================================================================
Tuberculosis (TBC) is an infectious disease that remains a global health problem, including in Indonesia. East Java has a relatively high number of TBC cases, thus requiring an in-depth analysis of the factors influencing its spread. Previous studies have applied spatial models such as Geographically Weighted Poisson Regression (GWPR) and
Geographically Weighted Regression (GWR), but these were limited to global relationship analysis without considering more complex spatial distributions. This study aims to
compare the Geographically Weighted Generalized Poisson Regression (GWGPR) and Geographically Weighted Negative Binomial Regression (GWNBR) methods in analyzing
the factors affecting TBC spread by considering spatial aspects and overdispersion. The results show that in 2023, there were 85.060 TBC cases in East Java Province with an
average of 2238,421 cases per city/regency and a standard deviation of 1885,143, indicating a wide spread in the data. The GWGPR model resulted in four regional groups based
on significant variables, while GWNBR formed five groups. Based on the Mean Square Error (MSE) and Corrected Akaike Information Criterion (AICc), the GWGPR model performed better, with MSE of 976165,036 and AICc of 613,1458. The Breusch-Pagan test indicated spatial heterogeneity, while Moran’s I test showed no spatial autocorrelation, suggesting that the adaptive bisquare kernel spatial weighting is appropriate for both models. Therefore, the GWGPR method is deemed more suitable for modeling the number of TBC cases in cities/regencies of East Java than the GWNBR method.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Tuberculosis, Geographically Weighted Generalized Poisson Regression (GWGPR), Geographically Weighted Negative Binomial Regression (GWNBR), Tuberculosis, Geographically Weighted Generalized Poisson Regression (GWGPR), Geographically Weighted Negative Binomial Regression (GWNBR) |
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis |
| Depositing User: | Galuh Putri Nabila |
| Date Deposited: | 31 Jul 2025 09:30 |
| Last Modified: | 31 Jul 2025 09:30 |
| URI: | http://repository.its.ac.id/id/eprint/125110 |
Actions (login required)
 |
View Item |