Figure 3.

This heat-map highlights the similarity of the communities discovered by means of the two adopted algorithms (i.e., FNCA and LPA) applied to theuniformsample. Almost the totality of communities discovered share a high fraction of members (in average the 91%), according the Jaccard similarity computed pairwisely selecting the most similar communities in the partitions.

Ferrara EPJ Data Science 2012 1:9   doi:10.1140/epjds9
Download authors' original image