Fig. 12From: Efficiency of random swap clusteringObserved probability (%) of failure (q) for S1–S4 (left) and for the high dimensional data sets (right). The observed values are plotted, and the theoretical estimates are shown as blueBack to article page