Fig. 5.

Comparing the prediction performance of DTi2Vec and state-of-the-art methods in random CV setting (in terms of AUPR with standard errors are shown) using the Yamanishi_08 and FDA_DrugBank datasets