Skip to main content

Table 14 Macro-AUPR obtained by focusing on Organ abnormality

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

Method[1-10][11-30][31-100][101-300][ ≥301]
NMF-Organ0.0300.1900.3550.4780.446
NMF-PPN-Organ0.0330.2050.3710.4950.486
NMF-NHPO-Organ0.0330.2070.3690.4900.485
NMF-All0.0310.1930.3630.4770.449
NMF-PPN-All0.0320.2040.3700.4860.460
NMF-NHPO-All0.0320.2090.3730.4820.462
HPOAnnotator-All0.0350.2120.3740.4930.485
  1. The three rows with “Organ" use only organ abnormality for training, while the other four rows with “All" take all sub-ontologies for training.
  2. Method performs best in terms of this evaluation metric are in boldface