The stability of a good clustering
If we have found a "good" clustering C of data set X, can we prove that C is not far from the (unknown) best clustering C* of this data set? Perhaps surprisingly, the answer to this question is sometimes yes. We can show bounds on the distance( C, C* ) for two clustering cost functions: the Normalized Cut and the squared distance cost of K-means clustering. These bounds exist in the case when the data X admits a "good" clustering for the given cost.
Author: Marina Meila, Department Of Statistics, University Of Washington