Figure 1
Two binary trees connected at one node.
The x-axis shows the number of nodes in community B as a fraction f of nodes in community A. The triangles and squares show the performance of the two operators in detecting communities as measured by the normalised mutual information (NMI): 0 ≤ NMI ≤ 1, where NMI = 1 means perfect community detection and NMI = 0 means random allocation of nodes to communities (see Methods for more details). (a) 400 nodes in community A. Number of nodes in community B varies from from 40 to 400. (b) 500 nodes in community A. Number of nodes in community B varies from 50 to 500.
