Some issues in robust clustering

Cluster analysis is about finding groups in data. Robust statistics is about methods that are not affected strongly by deviations from the statistical model assumptions or moderate changes in a data set. Particular attention has been paid in the robustness literature to the effect of outliers. Outliers and other model deviations can have a strong effect on cluster analysis methods as well. There is now much work on robust cluster analysis, see [1, 19, 9] for overviews.