ANALISIS REGRESI BAYES LINEAR SEDERHANA DENGAN PRIOR NONINFORMATIF

Abstract

The aim of this study is to apply Bayesian simple linear regression using noninformative prior. The data used in this study is 30 observational data with error generated from normal distribution. The noninformative prior was formed using Jeffreys’ rule. Computation was done using the Gibbs Sampler algorithm with 10.000 iteration. We obtain the following estimates for the parameters, with 95% Bayesian confidence interval (0,775775; 2,626025), with 95% Bayesian confidence interval (2,948; 3,052), and with 95% Bayesian confidence interval (0,375295; 1,114). These values are not very different compared to the actual value of the parameters.