An ensemble learning approach for improving drug–target interactions prediction

Abstract

Prediction of drug–target interactions (DTIs) plays an important role in the drug discovery process. Unfortunately, experimental determination of interactions between drug compounds and target proteins remains a challenging task. Current available heterogeneous data motivate us to develop an effective data integration approach to exploit intrinsic correlations of known interactions between drugs and targets and to predict new interactions. We propose an ensemble learning approach to integrate previously developed methods and to improve the prediction performance. In particular, our algorithm employs a stacking framework, which uses a support vector machine (SVM) classifier as the meta learner to achieve better prediction results.