Table 7.

Robustness evaluation of our classifiers

Data set	Accuracy	AUC	F1-score $\oplus$
$DIL{I}^{\mathit{shuffled}}$	0.52	0.49	0.36
$DIL{I}^{random{\ominus }_1}$	0.92	0.91	0.89
$DIL{I}^{random{\ominus }_2}$	0.92	0.91	0.89
$DIL{I}^{random{\ominus }_3}$	0.93	0.92	0.91
$DIL{I}^{random{\ominus }_4}$	0.93	0.92	0.90
$DIL{I}^{random{\ominus }_5}$	0.92	0.91	0.90
$SCA{R}^{\mathit{shuffled}}$	0.63	0.51	0.26
$SCA{R}^{random{\ominus }_1}$	0.93	0.89	0.86
$SCA{R}^{random{\ominus }_2}$	0.94	0.90	0.88
$SCA{R}^{random{\ominus }_3}$	0.93	0.90	0.86
$SCA{R}^{random{\ominus }_4}$	0.92	0.89	0.85
$SCA{R}^{random{\ominus }_5}$	0.93	0.89	0.86

${·}^{\mathit{shuffled}}$ corresponds to an experiment where class labels (i.e., $\oplus$ or $\ominus$) are randomly affected to drugs. ${·}^{random{\ominus }_i}$ correspond to experiments where negative examples (i.e., $\ominus$) are replaced by drugs randomly picked in the knowledge graph. Indices i from 1 to 5 refer to 5 different draws