Table 2 summarizes three cluster validity values of all seeds initialization algorithms. For fair comparison among algorithms, the number of clusters for each data set was set to the same as the number of classes. k of the CSI was set
to 5 while ρ of SCS was set to the average distance between instances