Skip to main content

Table 2 The prediction performances in 5CV,10CV and denovo validation, the best results are in the bold face

From: DDIGIP: predicting drug-drug interactions based on Gaussian interaction profile kernels

 The prediction performances(AUC)   
MethodFeature5CV10CVDenovo
WAEChemical data, biological data, phenotypic data0.95020.95300.9073
L1EChemical data, biological data, phenotypic data0.95700.9599
L2EChemical data, biological data, phenotypic data0.95610.9594
LPDrug-sub0.93560.93590.8993
 Drug-Label0.93640.93680.8994
 Drug-Off Label0.93740.93780.8997
DDIGIPChemical data, biological data, phenotypic data0.96000.96360.9262
  1. The represents that we did not compute the prediction performance because the prediction limit for new drugs.