Geographically Weighted Panel Regression (GWPR) and Geographically and Temporally Weighted Regression (GTWR) Methods in the Influence of Factors Affecting the Minimum Wage of Each Province in Indonesia
Apriani Monica Renta,Herawati Netti,Sutrisno Agus,Widiarti
DOI: https://doi.org/10.31305/rrijm.2024.v09.n07.015
2024-07-15
RESEARCH REVIEW International Journal of Multidisciplinary
Abstract:The study compares the Geographically Weighted Panel Regression (GWPR) method and the Geographically Temporal Weighted Regression (GTWR) method using a spatial and temporal approach in the data. Both of these methods are developments by one of the Geographically Weighted Regression (GWR) methods. GWPR, which is a combination of the GWR model with panel data regression to analyze spatial variations between regions on the influence of the provincial minimum wage determination factor in Indonesia. Meanwhile, GTWR is used to analyze the presence of non-stationarity in its data if there is spatial and temporal simultaneously. In this study, both methods, namely GWPR and GTWR, will show variations in the influence of variables in each location by showing their local parameters with kernel weighting. In selecting the optimum bandwidth of the kernel weight function using cross validation (CV) with the minimum value to be used. After that, when measuring the best model of the two methods, which is GWPR and GTWR, it will be shown by the largest R2 value, RMSE and, AIC with the smallest value. From the results of the study between GWPR and GTWR, namely the GTWR method has an R2 value (93.73%) and the GWPR method has an R2 value (67.32%). Therefore, the R2 value of the GTWR method is greater than the GWPR method. The results of the study indicate that the GTWR method is a better method in handling spatial-temporal.