The minimum sum of absolute errors regression: A robust alternative to theleast squares regression

This paper concerns the minimum sum of absolute errors regression. It is a more robust alternative to the popular least squares regression whenever th ere are outliers in the values of the response variable, or the errors foll ow a long tailed distribution, or the loss function is proportional to the absolute errors rather than their squared values. We use data from a study of interstitial lung disease to illustrate the method, interpret the findin gs, and contrast with least squares regression. We point out some of the pr oblems with the least squares analysis and show how to avoid these with the minimum sum of absolute errors analysis. (C) 1999 John Wiley & Sons, Ltd.