“…The superparamagnetic clustering method, as it was dubbed in the original paper, has been successfully applied to different research problems [Agrawal and Domany, 2003;Domany, 2003;Domany et al, 1999;Einav et al, 2005;Getz et al, 2000a;Kullmann et al, 2000;Ott et al, 2004Ott et al, , 2005Quiroga et al, 2004] and various extensions of the method were proposed, including the iterative coupled two-way clustering [Getz et al, 2000b;Getz and Domany, 2003], the modified Hamiltonian and an improved version of the cluster update algorithm for the image segmentation problem [von Ferber and Woergoetter, 2000], a sequential extension for inhomogeneous clusters [Ott et al, 2004[Ott et al, , 2005, and a spin-glass Hamiltonian for complex networks data [Reichardt and Bornholdt, 2004]. Murua et al (On Potts model clustering, kernel K-means and density estimation, J Comput Graph Stat, under review) related the Potts model based clustering to kernel-based clustering methods and density estimation and introduced yet another penalized Hamiltonian to favor clusters of equal sizes.…”