Table 4

SVM classification using different data sources
Data source AUC
AmiGO 0.886
BioCyc 0.807
CDD 0.876
GenNav 0.940
InterPro 0.883
Kegg 0.795
Kegg (pathways) 0.815
Pdb 0.875
TigrFam 0.872

Above are results by source, when empty graphs (query graphs with no returned results) are excluded from training and testing; the scores are thus those of each source given data from that source was available.

Cadag et al.

Cadag et al. BMC Bioinformatics 2012 13:321   doi:10.1186/1471-2105-13-321

Open Data