Evaluation methods for clustering based on the distribution of examples.