ANALISIS BAYESIAN UNTUK REGRESI KUANTIL DENGAN MENGGUNAKAN ALGORITMA GIBBS SAMPLING

• Main Authors: , ANNISA HANIF, , Prof. Dr. Subanar, Ph.D
• 格式: Theses and Dissertations NonPeerReviewed
• 出版: [Yogyakarta] : Universitas Gadjah Mada 2013
• 主題: ETD
• 在線閱讀: https://repository.ugm.ac.id/123628/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=63742

Quantile regression has received increasing attention both from a theoretical and from an empirical view point. It is a statistical procedure that minimizing sums of asymmetrically weighted absolute and can be used to explore the relationship between quantile of response distribution. Quantile regression can be used to overcome the limitation of linear regression to analyze data not symmetric and quantile regression is useful if the distribution of datais not homogeneous. Quantile regression can be estimated using Bayesian method. Bayesian method is a method of analysis based on information from sample and prior information. Combination of those informations is called posterior. For looking posterior dsitribution often result in calculation can not be solved with analytical so Gibbs sampling approach is used. Estimation of parameters in the model is the mean of the posterior distribution that obtained from Gibbs sampling process. In this paper discused quantile regression using an asymmetric Laplace distribution from Bayesian point of view. Gibbs sampling is used to find the estimator of quantile regression model based on a location-scale mixture representation asymmetric Laplace distribution. The case study in this paper discusses the factors that effect the gold price. The estimation result of quantile regression using Bayesian method will be compared with linear regression using OLS method, and compared with quantile regression method. And then it was concluded that the Bayesian method is better than the otherconclusion