Artikel Jurnal ESTIMASI PARAMETER MODEL REGRESI LINIER SEDERHANA BAYES DENGAN DISTRIBUSI PRIOR NONINFORMATIF JEFFREY
Download Jurnal Disinik
Abstract: The main objective in modeling observational data with regression models is to estimate the parameters. The parameter estimation can be prediction of point and interval parameters. The estimation of point parameters can be done by two methods, first referred to as the classical approach ( frequentist ). One of the techniques used in the classical method is the maximum method likelihood . The second approach, known as the bayesian . Bayesian in addition to utilizing the inference process in the sample data taken from the population, also considers the initial distribution (distribution prior ). In the Bayes method the parameter is considered as a variable that describes the initial knowledge about the parameters before the observation is performed and expressed in a distribution called the prior distribution . After the observations were made, the information in the distribution prior was combined with information with sample data via Bayes's theorem, and the results were expressed as posterior distributions . In general the choice of distribution prior is done on the basis of whether or not information about parameters is known. In this article we discuss the estimation of the parameters of the linear regression model on the distribution pattern prior ie if information about the parameters is not available, prior noninformative Jeffrey is used. Prior This noninformative does not have a significant influence on the distribution so that the information obtained from observational data is more objective.
Keywords: Bayesian Regression, Parameter Estimation, Prior Distribution, Noninformative Priority