From: A prediction-based resampling method for estimating the number of clusters in a dataset
Clest parameter | Value |
---|---|
Maximum number of clusters | M = 10 for microarray data |
 | M = 5 for simulated data |
Number of learning/test set iterations | B = 20 |
Number of reference datasets | B0 = 20 |
Size of learning sets b | 2n/3 |
Clustering procedure | PAM |
Classifier | Linear discriminant analysis with diagonal covariance matrix - DLDA |
Reference null distribution | Uniformity hypothesis |
External index | Fowlkes and Mallows [27] external index, FM |
Maximum p-value | p max = 0.05 |
Minimum difference statistic | d min = 0.05 |