Penerapan Regresi Spasial pada Data Kemiskinan Kabupaten/Kota di Provinsi Jawa Barat Tahun 2012

Abstract

Poverty is a social problem that normally occurs in the middle of society, especially in developing countries. One important aspect to support poverty reduction strategy is the availability of accurate poverty data. Good data can be used to evaluate government policies on poverty, compare poverty between regions and determine the targeted destitute population with an am to improve their living conditions. One effort to overcome poverty is to identify the variables that affect it. In this research, poverty issue is examined by percentage of poor people districts/cities in West Java Province in 2012, making use of spatial regression analysis. The analysis showed that the spatial lag models (SAR) produced better allegation parameters than classic regression models in case of spatial correlation. This can be seen from the smaller AIC of SAR models (129.62) than classic regression models (133.06). In SAR models, variables that significantly affect the percentage of poor people in district/cities are the percentage of households using clean water (X10) and households with floor area per capita (X21).