Fig. 15From: Efficiency of random swap clusteringNumber of iterations needed to find the correct clustering for subsets of BIRCH2 when varying the size of data (left), and number of clusters (right). The black line is the average of 100 runs, and the 25 and 75% quartiles show the level of variation between the individual runs. The blue lines are linear (left) and quadratic (right) fits to the dataBack to article page