| dc.description.abstract | The Open Unemployment Rate is a critical indicator in evaluating the labor market conditions of a region. In North Sumatra Province, the heterogenity among districts suggests the presence of spatial variability in the determinants. This study aims to determine the Geographically Weighted Regression (GWR) modeling in the case of the Open Unemployment Rate in North Sumatra Province. This research adopts a quantitative method with a cross-sectional design. The analysis utilizes secondary data from 33 districts/cities in North Sumatra Province, involving three independent variables: population growth rate, labor force participation rate, poverty rate. Initial modeling was conducted using multiple linear regression (OLS) as a global benchmark, followed by spatial modeling using. Classical regression diagnostics and spatial tests, including Global Moran’s I and the Breusch-Pagan test, were employed to assess model assumptions and spatial structure. The findings reveal that, jointly, the three independent variables significantly affect the open unemployment rate; however, only labor force participation rate shows a statistically significant individual effect (p < 0.05). The GWR model yields a higher coefficient of determination (R² = 0.683) compared to the OLS model (R² = 0.663), indicating a marginal improvement in explanatory power. Nevertheless, the GWR model records a higher Akaike Information Criterion (AIC = 124.22) than the OLS model (AIC = 123.79), suggesting lower model efficiency. The Global Moran’s I statistic confirms significant positive spatial autocorrelation, while the Breusch-Pagan test indicates no spatial heterogeneity. In conclusion, while the GWR model captures local variations in the relationship between explanatory variables and the open unemployment rate, it does not substantially outperform the global model in terms of statistical efficiency. Nonetheless, GWR remains a valuable tool for exploring spatial dynamics in regional labor market analysis. | en_US |
