Proceedings of the 2nd International Seminar of Science and Applied Technology (ISSAT 2021)

Application of the Geographically Weighted Regression (GWR) with the Bi-Square Weighting Function on the Poverty Model in the City/Regency of West Java

Authors
Euis Sartika1, *, Sri Murniati2
1Departement of Commerce, Politeknik Negeri Bandung
2Departement of Refrigeration and Air Conditioning Engineering, Politeknik Negeri Bandung
*Corresponding author. Email: euis.sartika@polban.ac.id
Corresponding Author
Euis Sartika
Available Online 23 November 2021.
DOI
10.2991/aer.k.211106.031How to use a DOI?
Keywords
GWR; poverty; Bi-Square; coefficient of determination; AIC
Abstract

The Geographically Weighted Regression (GWR) analysis is considered the most appropriate analysis to describe the Poverty model, including location. By employing the GWR analysis, this study is aimed to find out the proper model of poverty in West Java Province, which can describe the geographical characteristics of the location or district/city in West Java with the Kernel Bi-Square Weighting function. The secondary data were taken from 2018 which consist of the response variable of the Poor Percentage (PP) and the independent variables which cover the Open Unemployment Rate (OUR), Human Development Index (HDI), Gross Regional Domestic Product (GRDP), Population Density Level (PDL), Regional Minimum Wage (RMW), Poor Population Percentage aged 15 years and high school (PPHS), and Literacy Rates (LR). These variables are estimated to influence the rate of poverty. Multiple regression (Global) and GWR regression (Local) were applied in the analysis, and the weighting function used was Bi-Square. Whereas the best model was selected using the criteria of the coefficient of determination R2 and the value of AIC. The results showed that the local GWR regression model has a coefficient of determination (R2) of 0.9253, meaning that the independent variables could explain 92.53% of the variation in the Poverty Percentage model. The remaining 7.47% is explained by other factors. Besides, the value of the global regression coefficient of determination is 0.7084. The AIC value for GWR is 352.437, and the AIC value for global regression is 363.227, meaning that the error value for GWR is smaller than the global regression. Thus, it can be concluded that the GWR (local) regression model is considered a better model. The variables that affect the percentage of poor people in the global regression model are the Open Unemployment Rate and the Regional Minimum Wage. Meanwhile, the variables that affect the percentage of poor people for 27 cities/districts of West Java vary.

Copyright
© 2021 The Authors. Published by Atlantis Press International B.V.
Open Access
This is an open access article under the CC BY-NC license.

Download article (PDF)

Volume Title
Proceedings of the 2nd International Seminar of Science and Applied Technology (ISSAT 2021)
Series
Advances in Engineering Research
Publication Date
23 November 2021
ISBN
10.2991/aer.k.211106.031
ISSN
2352-5401
DOI
10.2991/aer.k.211106.031How to use a DOI?
Copyright
© 2021 The Authors. Published by Atlantis Press International B.V.
Open Access
This is an open access article under the CC BY-NC license.

Cite this article

TY  - CONF
AU  - Euis Sartika
AU  - Sri Murniati
PY  - 2021
DA  - 2021/11/23
TI  - Application of the Geographically Weighted Regression (GWR) with the Bi-Square Weighting Function on the Poverty Model in the City/Regency of West Java
BT  - Proceedings of the 2nd International Seminar of Science and Applied Technology (ISSAT 2021)
PB  - Atlantis Press
SP  - 201
EP  - 207
SN  - 2352-5401
UR  - https://doi.org/10.2991/aer.k.211106.031
DO  - 10.2991/aer.k.211106.031
ID  - Sartika2021
ER  -