Figure 3
From: Supervised maximum-likelihood weighting of composite protein networks for complex prediction
Precision-recall AUC for yeast complex prediction, using the five weighting approaches for each of the six clustering algorithms and the COMBINED clustering strategy, for k = 10000 (lighter shade), k = 20000 (darker shade), and k = all edges (only for NOWEI). For CMC, MCL and HACO, three sets of clustering parameters are tried. The AUC for match_thres = 0.5 and match_thres = 0.75 are shown in each bar. SWC achieves highest precision-recall AUC for all clustering algorithms except IPCA and HACO, where it performs about evenly with STR. The COMBINED strategy achieves higher AUC compared to using any single clustering algorithm alone.