Regresi Logistik dengan Metode Bayes untuk Pemodelan Indeks Pembangunan Manusia Kabupaten/Kota di Pulau Kalimantan

Abstract

Human Development Index (HDI) is an indicator that can measure success in efforts to build the quality of human life. HDI is also a measure of the prosperity of a region which is observed based on three dimensions, namely health, education and economy. Based on HDI publication by the Central Statistics Agency in 2018, it showed that the scores of HDI for 56 districts/cities in Kalimantan Island only has two categories of  HDI  which are medium and high. Bayesian method is a parameter estimation technique that combines the likelihood and prior distribution functions. The estimation with Bayesian method was solved using Markov Chain Monte Carlo simulation (MCMC) with Gibbs Sampler algorithm. The aim of this study is  to examine the modelling of the factors that influence the HDI of districts/cities in Kalimantan Island and determine the accuracy of the model classification using logistic regression with Bayesian method. The data used is the HDI of districts/cities in Kalimantan Island in 2018. Bayesian method is a parameter estimation technique that combines the likelihood and prior distribution functions. The estimation with Bayesian method was solved using Markov Chain Monte Carlo simulation (MCMC) with Gibbs Sampler algorithm. The results of modelling and analysis on districts/cities HDI data on Kalimantan Island showed that the factors that significantly influence HDI are the number of paramedic, the number of health facility and the participation rate of high school. The results of the classification accuracy of the model amounted to 82,14% which resulted in 37 districts/cities are categorized as the HDI medium category and 19 districts/cities are categorized as the HDI high category.