Abstract
Objects can be clustered in many different ways. As a matter of fact there are several cluster analysis methods that can produce different clusterings on the same dataset. Moreover, even when a single algorithm is used, different alternative clusterings can easily be generated, simply by changing the initial conditions of the algorithm. This work proposes a flexible criterion based on copula function for comparing two partitions (or clusterings) of the same dataset. This criterion also allows measuring the amount of information lost and gained in changing from cluster C to clustering C'.