The receiver operating characteristic (ROC) curves of the three different methods to predict new known drugs for a given target on the four benchmark data sets by simulation 30 times of 10-fold cross validation test, (a) enzymes, (b) ion channels, (c) GPCRs and (d) nuclear receptors, drug-based similarity inference (DBSI): dot dash curve, target-based similarity inference (TBSI): solid curve, network-based inference (NBI): dash curve, FPR: false positive rate and TPR: true positive rate.

<p>The receiver operating characteristic (ROC) curves of the three different methods to predict new known drugs for a given target on the four benchmark data sets by simulation 30 times of 10-fold cross validation test, (a) enzymes, (b) ion channels, (c) GPCRs and (d) nuclear receptors, drug-based similarity inference (DBSI): dot dash curve, target-based similarity inference (TBSI): solid curve, network-based inference (NBI): dash curve, FPR: false positive rate and TPR: true positive rate.