On the distribution equivalence in cluster analysis