Figure 6
From: Discovering multi–level structures in bio-molecular data through the Bernstein inequality
Lymphoma data set. Empirical cumulative distribution functions of the similarity measures for different number of clusters k. Perturbation technique: random projections