Farida, Ira (2016) MODEL GEOGRAPHICALLY WEIGHTED REGRESSION (GWR) DENGAN PEMBOBOT KERNEL BISQUARE (Studi Kasus : Indeks Pembangunan Manusia (IPM) Tahun 2013 di Kabupaten dan Kota Provinsi Jawa Barat). S1 thesis, Universitas Pendidikan Indonesia.
|
Text
S_MAT_1202569_Title.pdf Download (163kB) | Preview |
|
|
Text
S_MAT_1202569_Abstract.pdf Download (259kB) | Preview |
|
|
Text
S_MAT_1202569_Table_of_content.pdf Download (337kB) | Preview |
|
|
Text
S_MAT_1202569_Chapter1.pdf Download (227kB) | Preview |
|
![[img]](http://repository.upi.edu/style/images/fileicons/text.png) |
Text
S_MAT_1202569_Chapter2.pdf Restricted to Staf Perpustakaan Download (403kB) |
|
|
Text
S_MAT_1202569_Chapter3.pdf Download (401kB) | Preview |
|
|
Text
S_MAT_1202569_Chapter4.pdf Restricted to Staf Perpustakaan Download (509kB) |
|
|
Text
S_MAT_1202569_Chapter5.pdf Download (259kB) | Preview |
|
|
Text
S_MAT_1202569_Bibliography.pdf Download (225kB) | Preview |
|
|
Text
S_MAT_1202569_Appendix1.pdf Restricted to Staf Perpustakaan Download (545kB) |
|
|
Text
S_MAT_1202569_Appendix2.pdf Restricted to Staf Perpustakaan Download (148kB) |
|
|
Text
S_MAT_1202569_Appendix3.pdf Restricted to Staf Perpustakaan Download (320kB) |
|
|
Text
S_MAT_1202569_Appendix4.pdf Restricted to Staf Perpustakaan Download (435kB) |
Abstract
Geographically Weighted Regression (GWR) merupakan regresi dengan memperhatikan aspek lokasi geografis. GWR adalah bentuk lokal dari regresi linier. Model GWR digunkan untuk mengetahui hubungan variabel prediktor dengan variabel respon, yang mana koefisien regresinya bersifat lokal untuk setiap lokasi pengamatan. Pada model GWR ini fungsi pembobot yang digunakan untuk penaksiran parameternya adalah menggunakan kernel bisquare, yang berdasarkan bandwidth-nya, kernel bisquare terdiri dari fixed bisquare dan adaptive bisquare. Tujuan penelitian ini adalah untuk mengetahui pembobotan yang lebih baik antara fixed bisquare dan adaptive bisquare dalam pemodelan Indeks Pembangunan Manusia (IPM) di kabupaten/kota Provinsi Jawa Barat Tahun 2013 dengan menggunakan GWR, serta untuk mengetahui faktor-faktor yang mempengaruhi IPM berdasarkan pembobotan terbaiknya. Setelah dilakukan pengujian terdapat heterogenitas spasial dalam kasus IPM di 17 kabupaten dan 9 kota Provinsi Jawa Barat, sehingga dapat dilanjutkan pemodelan menggunakan GWR. Berdasarkan hasil analisis, diperoleh bahwa pembobotan terbaik adalah fixed bisquare dengan nilai koefisien determinasi R^2= 0,994240 lebih besar dari pembobot dengan adaptive bisquare R^2= 0,984088 dan nilai kuadrat residual JK(S)= 3,809 yang lebih kecil dibandingkan pembobot dengan adaptive bisquare JK(S)= 10,524. Sehingga diperoleh faktor-faktor yang berpengaruh signifikan terhadap IPM dengan pembobot fixed bisquare adalah persentase Angka Harapan Hidup (AHH), persentase penduduk miskin, persentase Angka Partisipasi Sekolah (APS) usia 16-18 tahun, Produk Domestik Regional Bruto (PDRB) per kapita, persentase Tingkat Partisipasi Angkatan Kerja (TPAK), persentase penduduk dengan akses air minum layak, persentase Perilaku Hidup Bersih dan Sehat (PHBS), angka Expected Years of Schooling (EYS), dan persentase laju pertumbuhan ekonomi dengan model GWR yang berbeda-beda di setiap kabupaten kota Provinsi Jawa Barat. Kata Kunci : Indeks Pembangunan Manusia (IPM), Geographically Weighted Regression (GWR), Fixed Bisquare, Adaptive Bisquare. Geographically Weighted Regression (GWR) is a regression based on the aspect of geographic location. GWR is a local form of linear regression. GWR models is used to determine the relationship predictor variables with the response variable, which is local regression coefficient for each observation location. In this GWR models weighting function is used to estimate parameters by using bisquare kernel, based on its bandwidth, as bisquare kernel consists of fixed bisquare and adaptive bisquare. The purpose of this study was to determine the better weighting between fixed bisquare and adaptive bisquare in modeling the Human Development Index (HDI) in the district/city of West Java province in 2013 using a GWR, as well as to determine the factors that affect the HDI based on the weighted best. After testing it is found that there is spatial heterogeneity in the case of HDI in 17 counties and nine cities of West Java province, so it can modeling using GWR. Can be further carried out based on the analysis, the best weighting is fixed bisquare with coefficient of determination R^2 = 0.994240 higher than the adaptive bisquare weighting R^2= 0.984088 and squared residual value JK(S)= 3.809 lower than the adaptive bisquare weighting JK(S)= 10.524. It is also found that the factors that influence significantly of HDI with the fixed bisquare weighting is the percentage of life expectancy, the percentage of poor population, the percentage of school enrollment rate age 16-18, Gross Domestic Regional Product per capita, the percentage of Labour Force Participation Rate, the percentage of the population with access to improved drinking water, the percentage of PHBS, figures Expected Years of Schooling (EYS), and the percentage of economic growth rate with models of GWR are different in each district of West Java Province. Keywords : Human Development Index (HDI), Geographically Weighted Regression (GWR), Fixed Bisquare, Adaptive Bisquare
| Item Type: | Thesis (S1) |
|---|---|
| Additional Information: | No. Panggil: S_MAT_FAR_m-2016; Pembimbing : I. Nar Herrhyanto, II. H. Asep Syarif Hidayat |
| Uncontrolled Keywords: | Indeks Pembangunan Manusia (IPM), Geographically Weighted Regression (GWR), Fixed Bisquare, Adaptive Bisquare. |
| Subjects: | H Social Sciences > H Social Sciences (General) L Education > L Education (General) Q Science > QA Mathematics |
| Divisions: | Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Jurusan Pendidikan Matematika |
| Depositing User: | Mr. Cahya Mulyana |
| Date Deposited: | 14 Aug 2018 07:14 |
| Last Modified: | 14 Aug 2018 07:14 |
| URI: | http://repository.upi.edu/id/eprint/30558 |
Actions (login required)
 |
View Item |