Fig. 3
From: Prediction of problematic complexes from PPI networks: sparse, embedded, and small complexes
Precision-recall graphs for complex prediction in human. For clarity, a shows only the integrated approach (SWC+DECOMP+SSS), each of its constituent approaches, and the PPI+COMBINE approach, while b includes the individual clustering algorithms. c and d show the performance when the generated small clusters are removed, which lowers recall but increases precision. For human we use a matching threshold of l g_m a t c h=0.5 for large complexes, and s m_m a t c h=1 for small complexes