<![CDATA[Geographically Weighted Regression (GWR) is the most frequently used for spatial modeling. GWR produces local model parameter estimates for each observed point. Unfortunately, GWR is known to be numerically unstable and can produce extreme coefficient estimates. Spatially Clustered Regression (SCR) and Spatially Constrained Clusterwise Regression (SCCR) are new approaches that combine cluster identification and regression estimation in one stage. This research evaluates these approaches to develop poverty alleviation in East Java with the largest number of poor people in rural areas as per March 2023 according to BPS. The response variable used is the percentage of poor families. While the explanatory variables used are the percentage of female heads of households, the percentage of non-electricity families, the average years of schooling, the percentage of home ownership, and the percentage of agricultural laborers. The results of GWR and K-Means produced three clusters in East Java, SCR produced four clusters in East Java, and SCCR produced three clusters in East Java. Based on the AIC value, the best approach is SCR with a value of 1,614. Based on its grouping, SCR is better in forming cluster with adjacent locations rather than GWR + K-Means and SCCR. The variables that significant to the percentage of poor families are the percentage of agricultural laborers, the percentage of home ownership, and the percentage of female heads of households.]]>
Copyrights © 2025
