In addition to being faster than the ISNC algorithm, the SC-seed algorithm has the advantage that its starting seeds are not random--they are determined by the similarity coefficient algorithm.
The ISNC algorithm is significantly better than all other algorithms.
SC/ROC, MSC/ROC, ISNC and SC-Seed are not significantly different from each other,
ROC/ROC, SC/ROC, MSC/ROC, ISNC and SC-Seed are not sigificantly different from each other.
The ISNC algorithm was statistically better than all other algorithms or the primary measure (the grouping measure) and performed well on both secondary measures, for the problems in the randomly generated data set.
The computational requirements for the ISNC algorithm are higher than those for the other algorithms.