Bayesian Analysis of Student t Linear Regression with Unknown Change-Point and Application to Stock Data Analysis

This article devotes to studying the variance change-points problem in student t linear regression models. By exploiting the equivalence of the student t distribution and an appropriate scale mixture of normal distributions, a Bayesian approach combined with Gibbs sampling is developed to detect the single and multiple change points. Some simulation studies are performed to display the process of the detection and investigate the effects of the developed approach. Finally, for illustration, the Dow Jones index closed data of U.S. market are analyzed and three variance change-points are detected. Copyright Springer Science+Business Media, LLC. 2012