Indonesia Conference Directory

Abstract West Java in 2017 was the highest province with the highest number of unemployed people in Indonesia, amounting to 48 683.7 thousand people. This resulted in low income which triggered a low level of education and public health. Low education levels and high unemployment cause poverty. In this study information will be examined regarding what factors influence the unemployment rate in West Java for the period 2017. The various factors affecting Open Unemployment Warning include: population density, GRDP, Human Development Index (HDI), Labor Force Participation Rate (TPAK), and Regional Minimum Wage (UMR). This study includes the location (spatial) element in modeling and taking the object of the location of cities / districts in West Java, because geographically the cities / districts in West Java are quite varied and the model formed illustrates the characteristics of each region in West Java. Analysis that accommodates location (spatial) aspects is Geographically Weighted Regression (local regression). Other analyzes used are descriptive analysis and multiple linear regression analysis (global regression). Assumptions that must be met to model global regression, namely: normality of data, non-autocorrelation, homocedasticity, no multicollinearity, and the estimated parameters obtained are global in nature (Supranto, 2010). GWR regression analysis showed spatial heterogeneity (location). The dependent variable (Y) is the Open Unemployment Rate, the independent variables include: Population Density Level, GRDP, UMR), TPAK (Active Labor Participation Rate), and HDI (Human Development Index). The results showed that the unemployment model in West Java formed from the analysis of Geographically Weighted Regression (GWR) or local regression gave a coefficient of determination (R2) of 89.20%, meaning that 89.20% of the Open Unemployment Rate in West Java was influenced by variables Population Density Level, GRDP, UMR (Regional Minimum Wage), TPAK (Active Participation Work Rate), and HDI (Human Development Index). The remaining 11.80% is influenced by other factors. While multiple linear regression analysis (global regression) provides a coefficient of determination (R2) of 61.51% and the remaining 38.49% is influenced by other factors. The AIC (akaike information criterion) value of the gwr model is 89.730766 smaller than the global regression model that is equal to 98.791017. this shows that the local regression model (GWR) gives a smaller error value than the global regression model. in other words, the local regression model is relatively better than the global regression model. The independent variables that have a significant effect on the model vary for each local regression model. There are 27 different combinations of local regression models according to the number of cities / districts in West Java. Research output is a scientific article and unemployment model that can be used as a guide for relevant parties to take policy in order to reduce unemployment in West Java.