Logistic Discrimination Using Robust Estimators

24 Pages Posted: 27 Jan 2006

See all articles by Christophe Croux

Christophe Croux

Université Libre de Bruxelles (ULB) - European Center for Advanced Research in Economics and Statistics (ECARES); Catholic University of Leuven (KUL) - Department of Applied Economics

Gentiane Haesbroeck

University of Liège. Department of Mathematics

Kristel Joossens

KU Leuven - Faculty of Business and Economics (FEB)

Abstract

Logistic regression is frequently used for classifying observations into two groups. Unfortunately there are often outlying observations in a data set, who might affect the estimated model and the associated classification error rate. In this paper, the effect of observations in the training sample on the error rate is studied by computing influence functions. It turns out that the usual influence function vanishes, and that the use of second order influence functions is appropriate. It is shown that using robust estimators in logistic discrimination strongly reduces the effect of outliers on the classification error rate. Furthermore, the second order influence function can be used as diagnostic tool to pinpoint outlying observations.

Keywords: Classification, Diagnostics, Discrimination, Error rate, Influence function, Logistic regression, Robustness

Suggested Citation

Croux, Christophe and Haesbroeck, Gentiane and Joossens, Kristel, Logistic Discrimination Using Robust Estimators. Available at SSRN: https://ssrn.com/abstract=876912 or http://dx.doi.org/10.2139/ssrn.876912

Christophe Croux (Contact Author)

Université Libre de Bruxelles (ULB) - European Center for Advanced Research in Economics and Statistics (ECARES) ( email )

Ave. Franklin D Roosevelt, 50 - C.P. 114
Brussels, B-1050
Belgium

Catholic University of Leuven (KUL) - Department of Applied Economics ( email )

Leuven, B-3000
Belgium

Gentiane Haesbroeck

University of Liège. Department of Mathematics ( email )

B-4000 Liege
Belgium

Kristel Joossens

KU Leuven - Faculty of Business and Economics (FEB) ( email )

Naamsestraat 69
Leuven, B-3000
Belgium

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
140
Abstract Views
1,059
Rank
372,989
PlumX Metrics