REGRESI LOGISTIK MULTINOMIAL BAYESIAN DENGAN ALGORITMA GIBBS SAMPLING UNTUK MENENTUKAN FAKTOR-FAKTOR TINGKAT KEMISKINAN DI INDONESIA

Poverty is a state of deprivation experienced by individuals or groups with monthly per capita expenditure that is insufficient to meet basic needs. Based on Indonesia's poverty profile released by the Statistics Indonesia (BPS) in March 2024, it was recorded that 9.03% of Indonesia's population was declared poor, which is still far from the poverty reduction target of 6.5% to 7.5% targeted in the National Medium-Term Development Plan 2020-2024. One of the efforts that can be made to end poverty in Indonesia is to analyze what factors affect the poverty rate. The method used in this study is Bayesian multinomial logistic regression using the Markov Chain Monte Carlo (MCMC) Gibbs Sampling algorithm and the response variable used as a measure of poverty level is the poverty line which is an official indicator sourced from BPS. The results show that after 20,000 iterations, the Markov chain reaches a stationary state with the results of the credible interval test supported by the deviance test results stating that the factors that have a significant effect on the poverty rate in Indonesia in 2024 are GRDP at constant prices and average years of schooling.