Figure 10From: A prediction-based resampling method for estimating the number of clusters in a datasetCorrelation matrix, melanoma dataset. Images of the correlation matrix for the 31 melanoma mRNA samples based on expression profiles for (a) all p = 3,613 genes and (b) the p = 100 genes with the largest variance. The mRNA samples are ordered by class, as proposed in [30], first group B (blue), then group A (orange).Back to article page