Penerapan Multiscale Geographically Weighted Regression (MGWR) dalam Pemodelan Spasial Tingkat Pengangguran Terdidik di Provinsi Sumatera Utara

Sihombing, Sarah Tri Puji Florensia (2026) Penerapan Multiscale Geographically Weighted Regression (MGWR) dalam Pemodelan Spasial Tingkat Pengangguran Terdidik di Provinsi Sumatera Utara. Undergraduate thesis, UPN Veteran Jawa Timur.

[img] Text (Cover)
22083010020.-cover.pdf

Download (872kB)
[img] Text (Bab 1)
22083010020.-bab1.pdf

Download (345kB)
[img] Text (Bab 2)
22083010020.-bab2.pdf
Restricted to Repository staff only until 12 November 2028.

Download (843kB)
[img] Text (Bab 3)
22083010020.-bab3.pdf
Restricted to Repository staff only until 12 November 2028.

Download (506kB)
[img] Text (Bab 4)
22083010020.-bab4.pdf
Restricted to Repository staff only until 12 November 2028.

Download (2MB)
[img] Text (Bab 5)
22083010020.-bab5.pdf

Download (323kB)
[img] Text (Daftar pustaka)
22083010020.-daftarpustaka.pdf

Download (188kB)
[img] Text (Lampiran)
22083010020.-lampiran.pdf
Restricted to Repository staff only

Download (126kB)

Abstract

Educated unemployment is a labor market issue associated with the labor market’s capacity to absorb educated workers. This study aims to analyze the spatial patterns of educated unemployment, identify the relationships between socioeconomic factors and educated unemployment across regions, and compare the performance of Multiscale Geographically Weighted Regression (MGWR) and Ordinary Least Squares (OLS). This study uses secondary data from the Central Statistics Agency (BPS) for 2024, covering 33 regencies/cities in North Sumatera Province. The analysis employs Local Indicator of Spatial Association (LISA) and Multiscale Geographically Weighted Regression (MGWR), using a K-Nearest Neighbors (KNN) spatial weight matrix with five nearest neighbors. The results indicate significant positive spatial autocorrelation, with a Moran’s I value of 0.3002 and a p-value of 0.002. The LISA analysis identifies six regions with significant local spatial patterns, comprising three High-High regions, two Low-Low regions, and one High-Low region. The MGWR results show that GRDP per Capita has a statistically significant negative coefficient in nine regencies/cities, whereas the Regency/City has a statistically significant positive coefficient in one regency/city. Compared with OLS, MGWR produces a lower Residual Sum of Squares (RSS) and higher R^2 and 〖Adjusted R〗^2 values, although its AICc value is higher. The research findings are implemented in an interactive R Shiny dashboard to visualize spatial patterns and variations in relationships between variables, providing information to support the formulation of region-based employment policies.

Item Type: Thesis (Undergraduate)
Contributors:
ContributionContributorsNIDN/NIDKEmail
Thesis advisorDamaliana, Aviolla TerzaNIDN0002089402aviolla.terza.sada@upnjatim.ac.id
Thesis advisorSaputra, Wahyu Syaifullah JauharisNIDN0725088601wahyu.s.j.saputra.if@upnjatim.ac.id
Subjects: Q Science > QA Mathematics
Divisions: Faculty of Computer Science > Departemen of Data Science
Depositing User: Sarah Tri Puji Florensia Sihombing
Date Deposited: 13 Nov 2026 02:24
Last Modified: 13 Nov 2026 02:24
URI: https://repository.upnjatim.ac.id/id/eprint/61326

Actions (login required)

View Item View Item