MODEL REGRESI PANEL SPASIAL PADA PENDUDUK MISKIN KABUPATEN/KOTA DI JAWA TENGAH

Septiyaningrum. 2016. SPATIAL PANEL REGRESSION MODEL ON POOR PEOPLE OF REGENCY/CITIES IN CENTRAL JAVA. Faculty of Mathematics and Natural Sciences. Sebelas Maret University. Poverty is a complex problem that resulted in social inequality. World Bank stated that the poverty reduction in Indonesia is getting slow down. In 2014, BPS stated the highest percentage of poor people is in Java Island. Central Java placed the second highest rank after D. I. Yogyakarta. Poor people data used is a combination between cross section and time series data called panel data. If poor people in one region to another is influenced by geographical conditions then panel data has a spatial effect. The spatial effect is used to determine how big correlation of some factors geographically. The research is aimed to determine factors that influence poor people of regency/cities in Central Java using spatial panel regression. The spatial weighting use the rook contiguity weighting because each region touch on side to another region. The results of this implementation are that the factors which influence poor people of regency/cities in Central Java are literacy rate and the expenditure per capita with the best model which is obtained SEM random effect. Keywords: poor people, spatial panel regression, rook contiguity weighting, SEM random effec