Pemodelan Angka Harapan Hidup Di Papua Dengan Pendekatan Geographically Weighted Regression

Tanadjaja, Ardianto (2017) Pemodelan Angka Harapan Hidup Di Papua Dengan Pendekatan Geographically Weighted Regression. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1312100150-Undergraduate-Theses.pdf - Published Version

Download (4MB) | Preview

Abstract

Papua merupakan salah satu bagian dengan Angka Harapan Hidup (AHH) yang rendah di Indonesia bahkan nilai AHH di Papua sendiri berada di bawah nilai rata-rata AHH negara Indonesia. Berdasarkan fakta dan penelitian terdahulu, diketahui bahwa AHH dipengaruhi oleh faktor yang bervariasi disetiap daerah, sehingga digunakan pemodelan menggunakan metode Geographically Weighted Regression. Berdasarkan hasil analisis, pada aspek spasial data telah memenuhi asumsi heterogenitas spasial menggunakan uji Breusch Pagan dengan P-value sebesar 0.17 dan terdapat dependensi spasial menggunakan uji Moran’s I dengan P-value sebesar 0.178. Fungsi pembobot GWR yang digunakan adalah fungsi pembobot Adaptive Gaussian dengan CV minimum sebesar 5.20 serta bandwith sebesar 0.7. Terjadi peningkatan nilai koefisien determinasi (R2) menjadi 98.9 persen pada model GWR dan penurunan nilai Sum Square Error (SSE) menjadi 0.09. Faktor signifikan yang mempengaruhi AHH di setiap kabupaten/kota di Papua adalah persentase rumah tangga menggunakan sumber air minum layak, harapan lama sekolah, lama pemberian ASI serta rasio bidan per 10.000 penduduk. ====================================================================================== Papua is one of the regions in Indonesia with the low Life Expectancy number (AHH—Angka Harapan Hidup). Based on facts and prior researches, it is known that AHH is affected by factors which vary in every region, thus the modeling of GWR method is utilized. The assumed variables which affect AHH in Papua are the percentage of neighborhood with improved drinking water sources, the percentage of neighborhood with improved sanitation, the school duration expectation, and many more. Based on the analysis result, the data has fulfilled the heterogeneity spatial assumption with P-value of 0.172 and there was a spatial dependency of P-value of 0.178. The weighting function used was Adaptive Gaussian weighting with the bandwidth of 0.7. There was an increase of R2 value to 98.9 percent in GWR model and a decrease of SSE value to 0.09. Thus, in this case, the GWR model is better in modeling compared to OLS model. The percentage of neighborhood with improved drinking water sources, the school duration expectation, the breast-feeding duration, and the ratio of obstetrician are the significant cause of AHH in every city and district in Papua with significance level of 18 percent.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: AHH; Dependensi Spasial; Heterogenitas Spasial; GWR; R2; SSE; Poverty; Spline semiparametric regression; West Java Provinces
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA278.2 Regression Analysis
Divisions: Faculty of Mathematics and Science > Statistics > (S1) Undergraduate Theses
Depositing User: - ARDIANTO TANADJAJA
Date Deposited: 20 Apr 2017 02:59
Last Modified: 08 Mar 2019 07:06
URI: http://repository.its.ac.id/id/eprint/3777

Actions (login required)

View Item View Item