Figure 4: Clustering of simulated data (Figure 1) by Chameleon illustrating increasing rates of mis-classification with increasing neighbourhood size. Samples with the same colour were placed in the same cluster.