Khoiriyah, Usnidatul (2018) Pemodelan Geographically Weighted Ordinal Logistic Regression (GWOLR) dengan menggunakan Pembobot Fixed Gaussian Kernel. Undergraduate thesis, Universitas Islam Negeri Maulana Malik Ibrahim.
Text (Fulltext)
14610072.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (3MB) | Request a copy |
Abstract
ABSTRAK
Regresi logistik ordinal sering digunakan sebagai metode yang dapat mengetehui hubungan antara variabel respon yang memiliki skala nominal atau ordinal dengan satu atau lebih variabel prediktor yang bersifat kontinu atau kategorik. Pada regresi logistik, variabel respon yang memiliki skala ordinal dapat digolongkan ke dalam bentuk Geographically Weighted Ordinal Logistik Regression (GWOLR). Model GWOLR merupakan gabungan dari model GWR dan OLR (Ordinal Logistic Regression). Estimasi parameter model GWOLR menggunakan metode maksimum likelihood.
Tujuan penelitian ini adalah memodelkan GWOLR dengan menggunakan pembobot fixed gaussian kernel yang diaplikasikan pada data rata-rata Usia Kawin Pertama (UKP) Wanita di Provinsi Jawa Timur pada Tahun 2012. UKP di klasifikasikan menjadi empat kategori yang berskala ordinal yaitu child marriage, early marriage, marriage at maturnity, dan late marriage.Dengan variabel respon (Y) nya adalah rata-rata Umur Kawin Pertama wanita, persentase penduduk wanita yang tinggal di pedesaan (X_1), rata-rata lama sekolah wanita (X_2 ), persentase penduduk wanita berusia 15 tahun ke atas yang bekerja (X_3), persentase penduduk miskin (X_4), pendapatan perkapita (X_5), gini ratio (X_6), dan pembagian daerah tapal kuda (X_7). Hasil penelitian pemodelan spasial GWOLR dengan menggunkan fungsi pembobot fixed gaussian kernel adalah
g_1 (x)=logit [P ̂(Y≤1├|x┤)]=-7,981+3,146〖 X〗_1+7,582 X_2-1,378 X_4
g_2 (x)=logit [P ̂(Y≤2├|x┤)]=0,196+3,146〖 X〗_1+7,582 X_2-1,378 X_4
g_3 (x)=logit [P ̂(Y≤3├|x┤)]=8,993+3,146〖 X〗_1+7,582 X_2-1,378 X_4
Pemetaan atau pengelompokan kabupaten/kota berdasarkan pada kesamaan prediksi rata-rata UKP wanita. Persentase ketetapan klasifikasi rata-rata UKP wanita berdasarkan model GWOLR terbaik sebesar 84,211%.
ABSTRACT
An ordinal logistic regression is often used as a method that can determine the relationship between response variables that have a nominal or ordinal scale with one or more predictor variables that are continual or categorical. In logistic regression, response variables that have an ordinal scale can be classified into Geographically Weighted Ordinal Logistics Regression (GWOLR). It is a combination of GWR and OLR (Ordinal Logistic Regression) models. The parameter estimation of the GWOLR model uses the maximum likelihood method.
The purpose of this study is to examine the GWOLR pattern using a fixed gaussian kernel weight which is applied to the data of the average First Marriage Age (UKP) of women in East Java in 2012. UKP is classified into four ordinal scale categories; there are child marriage, early marriage, marriage at maturity, and late marriage. By using these following response variables(Y) the average of women's first marriage age, the percentage of women's population in the countryside(X_1), the average duration of women's education〖(X〗_2),the percentage of working women aged 15 years old and above(X_3 ),the percentage of poor people(X_4 ),income per capita(X_5 ), gini ratio (X_6 ),tapal kuda regional division(X_7 ).
The results of the GWR spatial modeling study using the fixed gaussian kernel weighting function are :
g_1 (x)=logit [P ̂(Y≤1├|x┤)]=-7,981+3,146〖 X〗_1+7,582 X_2-1,378 X_4
g_2 (x)=logit [P ̂(Y≤2├|x┤)]=0,196+3,146〖 X〗_1+7,582 X_2-1,378 X_4
g_3 (x)=logit [P ̂(Y≤3├|x┤)]=8,993+3,146〖 X〗_1+7,582 X_2-1,378 X_4
The city mapping or grouping is based on the similarity in predictions of women's UKP. Therefore, the precision percentage of the average classification of women's UKP based on the best GWR model is about 84.211%.
Item Type: | Thesis (Undergraduate) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Supervisor: | Harini, Sri and Rahman, Hairur | |||||||||
Contributors: |
|
|||||||||
Keywords: | Regresi Logistik; GWOLR; Maximum Likelihood Estimation; Gaussian Kernel; Usia Kawin Pertama Wanita Logistic Regression; GWOLR; Maximum Likelihood Estimation; Gaussian Kernel; Age of First Married Women (UKP) | |||||||||
Departement: | Fakultas Sains dan Teknologi > Jurusan Matematika | |||||||||
Depositing User: | Moch. Nanda Indra Lexmana | |||||||||
Date Deposited: | 17 Mar 2023 13:25 | |||||||||
Last Modified: | 17 Mar 2023 13:25 | |||||||||
URI: | http://etheses.uin-malang.ac.id/id/eprint/48578 |
Downloads
Downloads per month over past year
Actions (login required)
View Item |