Robust Statistics
The aim of this course is to acquire knowledge and insight in robust statistical methods, and to be able to apply those methods to real data.
Practical information:
Want to register?
- Prerequisites: familiarity with basic multivariate statistical methods.
- Price: €275-€1100
Organised by:
Robust statistical methods are resistant to outlying observations in the data, and hence are also able to detect these outliers. In this course we will introduce modern robust statistical methods for univariate and multivariate data.
We study several robust estimators of location, scale, skewness, correlation, covariance and regression. For the analysis of high-dimensional data we discuss robust estimators for principal component analysis, principal component regression and partial least squares regression. Finally we consider robust methods for classification. Also notions of robustness such as breakdown point and influence function will be introduced.
During the morning sessions we study the methods and their robustness properties. We also discuss computational issues. The afternoon sessions are devoted to the analysis of real data sets using Matlab or R software.
Teacher/speaker
Peter Rousseeuw
Peter Rousseeuw is professor in statistics at Katholieke Universiteit Leuven. He obtained his Ph.D. on the topic of robust statistics in 1981, and has been a full professor in universities in the Netherlands, Switzerland, and Belgium. He has (co-)authored over 160 papers and three Wiley-Interscience books.
In 2003 ISI-Thompson included him in their list of Highly Cited Mathematicians. He is an elected member of the International Statistical Institute and a fellow of the Institute of Mathematical Statistics and the American Statistical Association.
Related courses
ESANN 2026: European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning.
Symposium - Brugge - ESANN
