Skip to main content
. 2018 Jun 27;34(13):i52–i60. doi: 10.1093/bioinformatics/bty259

Table 3.

AUC values of the ROC curves for PPI interaction type prediction

Yeast
Human
Reaction Activation Binding Catalysis Reaction Activation Binding Catalysis
Resnik 0.5811 0.6023 0.5738 0.5792 0.5341 0.5331 0.5233 0.5810
Onto2Vec 0.5738 0.5988 0.5611 0.5814 0.5153 0.5104 0.5073 0.6012
Onto2Vec_LR 0.7103 0.7011 0.6819 0.6912 0.7091 0.6951 0.6722 0.6853
Onto2Vec_multiSVM 0.7462 0.7746 0.7311 0.7911 0.7351 0.7583 0.7117 0.7724
Onto2Vec_NN 0.7419 0.7737 0.7423 0.7811 0.7265 0.7568 0.7397 0.7713
Binary_GO_LR 0.6874 0.6611 0.6214 0.6433 0.6151 0.6533 0.6018 0.6189
Binary_GO_multiSVM 0.7455 0.7346 0.7173 0.7738 0.7246 0.7132 0.6821 0.7422
Binary_GO_NN 0.7131 0.6934 0.6741 0.6838 0.6895 0.6803 0.6431 0.6752

Note: The best AUC value for each action is shown in bold.