Fig. 12From: Overlapping community finding with noisy pairwise constraintsCritical difference plots from Friedman aligned rank test with Finner p value correction with significance level \(\alpha = 0.05\) comparing Hybrid (the best performing variant from the previous experiment), SLPA and AL-SLPA algorithms on small and large synthetic networks. Algorithms which are not connected with the horizontal dark line are significantly different than each other. Lower ranks indicate better overall performanceBack to article page