Bayesian Approach to identify Masking and Swamping Problems in the Multivariate-Multiple Regression Analysis

Abstract

The Bayesian method developed by Chaloner and Brant to see whether, for instance, the outlier observation is included in the linear model or not, focuses on posterior distributions of realized but unobserved errors [5]. This method has been expanded by Varbanov for the multivariate-multiple regression model in a way to provide analysis opportunities for separate observations in terms of their being an outlier [13]. This study expanded the proposed method by Varbanov to offer the opportunity for analyzing whether observations in groups are an outlier or not. Therefore, it is likely to claim the existence of masking and swamping problems.

 Key Words: Realized but unobserved errors, masking and swamping problems, outlier observation.

 

Keywords

• -