Introduction to Robust and Quasi-Robust Statistical Methods

The first part book surveys available methods of robust and quasi-robust statistical methods. The theory is presented in an expository style and in a unifying manner which allows seemingly disparate items to find their place in a common structure. It then becomes gradually clear that the techincal requirements leading to robustness are very demanding. The second part of the book treats the methods as they are encountered in real life situations. Robustness requirements are relaxed a little and "quasi-robust" estimators are obtained; the latter are much more reliable than the standard estimators without being as difficult to handle as the as the robust estimators. Algorithms are discribed and test cases are discsssed. This second part imore importans the t for the statistician who routinely processes data sets.