Essentially one defines clustering as a partitioning of arbitrary data objects into groups, such that the objects in each group, or cluster, have common traits, with respect to a similarity function.