Pangulimang, Joice (2016) Penaksiran Parameter Dan Pengujian Hipotesis Regresi Geographically Weighted Bivariate Zero-Inflated Poisson Studi Kasus : Jumlah Kasus Penderita HIV dan AIDS di Kabupaten Trenggalek dan Ponorogo 2012. Masters thesis, Institut Technology Sepuluh Nopember.
![[thumbnail of 1314201040- Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
1314201040- Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Metode statistik yang sering digunakan untuk menganalisis data count (cacahan) adalah regresi Poisson. Namun regresi Poisson tidak tepat jika dipakai dalam menganalisis data cacahan Zero-Inflated sehingga metode yang digunakan yaitu Zero-Inflated Poisson (ZIP). Pemodelan sepasang data cacahan berdistribusi Poisson dapat menggunakan Bivariate Zero-Inflated Poisson Regression (BZIPR). Pada penelitian sebelumnya mengaplikasikan model global BZIPR pada kasus jumlah penderita HIV dan AIDS. Padahal HIV dan AIDS merupakan penyakit yang mudah menular sehingga menyebar ke berbagai daerah (lokasi). Oleh karena itu, penelitian ini dikembangkan untuk model lokal yaitu Geographically Weighted Bivariate Zero-Inflated Poisson Regression (GWBZIPR) untuk mengetahui bentuk penaksiran parameter dan pengujian hipotesis serta mengetahui faktor-faktor yang berpengaruh terhadap HIV dan AIDS di Trenggalek dan Ponorogo tahun 2012. Variabel yang digunakan adalah jumlah kasus HIV, jumlah kasus AIDS, persentase kelompok umur, persentase tingkat pendidikan rendah (SMA), persentase jumlah tenaga kesehatan, persentase penduduk yang menggunakan kondom, persentase kegiatan penyuluhan, dan persentase jaminan kesehatan. Penaksiran parameter regresi GWBZIP dilakukan dengan metode Maximum Likelihood Estimation (MLE) menghasilkan bentuk non linear dan diselesaikan dengan iterasi Newton Raphson, sedangkan pengujian hipotesis dilakukan dengan metode Maximum Likelihood Ratio Test (MLRT) yang menghasilkan bentuk statistik uji mengikuti distribusi Chi-Square. Berdasarkan kesamaan variabel yang signifikan, terbentuk 7 kelompok untuk kasus HIV dan 6 kelompok untuk kasus AIDS. Sebagian besar kasus HIV di setiap Kecamatan dipengaruhi oleh persentase penduduk yang memiliki pengetahuan rendah (SMA), persentase penduduk yang memakai kondom, dan persentase kegiatan penyuluhan kesehatan. Sedangkan kasus AIDS dipengaruhi oleh persentase jamkesmas.
=============================================================================================================
Statistical methods are often used to analyze the data count is a Poisson regression. However, Poisson regression is not appropriate to be used in analyzing the data shredded Zero-inflated so that the method used is the Zero-inflated Poisson (ZIP). Modelling a pair of data can be shredded using bivariate Poisson distributed Zero-inflated Poisson Regression (BZIPR). In a previous study applying global models BZIPR in case the number of people living with HIV and AIDS. Though HIV and AIDS is a contagious disease that spread to different regions (locations). Therefore, this study was developed for the local model is Geographically Weighted Bivariate Zero-inflated Poisson Regression (GWBZIPR) to determine the shape parameter estimation and hypothesis testing and determine the factors that influence HIV and AIDS in Psychology and Ponorogo 2012. Variables used is the number of HIV cases, the number of AIDS cases, the percentage of the age group, the percentage of low education level (high school), percentage of health workers, the percentage of the population using condoms, the percentage of extension activities, and the percentage of health insurance. GWBZIP regression parameter estimation was conducted using Maximum Likelihood Estimation (MLE) to produce a non linear and solved by Newton Raphson iteration, whereas hypothesis testing was conducted using Maximum Likelihood Ratio Test (MLRT) which form the test statistic follows the Chi-Square distribution. Based on the similarity of the significant variables, formed seven groups for HIV cases and 6 groups for AIDS cases. Most cases of HIV in each district affected by the percentage of people who had low knowledge (SMA), the percentage of people who use condoms, and the percentage of health education activities. While the percentage of AIDS cases is influenced by the health card.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTSt 519.536 Pan p |
| Uncontrolled Keywords: | HIV/AIDS MLE, MLRT, Regresi Bivariat, Regresi BZIP, Regresi GWZIP, Regresi GWBZIP. |
| Subjects: | Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Mr. Tondo Indra Nyata |
| Date Deposited: | 20 Jan 2020 07:40 |
| Last Modified: | 20 Jan 2020 07:40 |
| URI: | http://repository.its.ac.id/id/eprint/72769 |
Actions (login required)
 |
View Item |