Simulation based validity testing of an outlier accommodation method for minimax absolute error regression

Abstract

In previous work, a mixed integer programming model was proposed for the systematic identification of potential outliers in connection with minimax absolute error regression. This method simultaneously estimates the regression parameters in the process of outlier identification. In this paper we report on some validity testing for the method based on simulated outlier inclusion. The results suggest that the method is effective in identification of such simulated outliers.