Skip to main content

Table 3 The results of the eight criteria obtained by 5 ×5-fold cross-validation over Data-201706 for the nine competing methods in total

From: HPOAnnotator: improving large-scale prediction of HPO annotations by low-rank approximation with HPO semantic similarities and multiple PPI networks

MethodAUCAUPRmicro-AUCmicro-AUPRmacro-AUCmacro-AUPRleaf-AUCleaf-AUPR
LR0.7750.0280.7600.0720.5790.0520.5320.020
BiRW0.8750.0660.8260.0960.7320.0560.5970.031
OGL0.7850.0510.7760.0780.6030.0340.5360.014
DLP0.9020.0730.8750.1000.7360.0940.6590.055
NMF0.9610.4960.9000.2730.7530.1390.7010.089
NMF-PPN0.9630.5250.9020.2810.7560.1420.7030.089
NMF-NHPO0.9650.5410.9030.2900.7560.1440.7020.094
AiPA0.9700.5590.9050.2950.7600.1460.7050.096
HPOAnnotator0.9710.5620.9070.2960.7600.1520.7060.097
  1. Method performs best in terms of this evaluation metric are in boldface