An Algorithm for Finding Best Matches in Logarithmic Expected Time
Distribution of the number of citations over years.