Pengembangan dan Aplikasi Geoinformatika Bayesian pada Data Kemiskinan di Indonesia (Studi Kasus Jawa Timur)

Abstract

Since a long time ago, poverty has been a problem that can not be solved. Indonesia’s Central Beureu of Statistics (CBS) survey on March 2011 show that there are 30.02 million people or 12.49% of total Indonesian are considered poor. From the point of view of many field of sciences the substance and the method to overcome this problem has become a very interesting topic of research. Based on statistical methods, poverty has become very interesting because there is an issue that there is an autocorrelation between data, spatial autocorrelation, error variance heterogenity, spatial interaction, and other statistical issues. The main objective of this research is to find factors that influence poverty rate in a region by developing spatial bayesian statistics. The methods developed in this research include Simultan Autoregressive (SAR), Conditional Autoregressive (CAR), Geographically Weighted Regression (GWR), Small Area Estimation (SAE) and hotspot detection. Based on the SAR Bayes model it is shown that the percented of people not graduating elementary school has a significant effect on poverty rate. While the increase of spatial autocorrelation will influence the poverty rate by 0.10 in East Java. Beside that by using hierarchical bayes logit normal model with nearest neighboor spatial weighted found that 40.93% of families of Jember is considered poor.