A modified version of the K-means algorithm with a distance based on cluster symmetry
Distribution of the number of citations over years.