Predicting drug target interactions using dimensionality reduction with ensemble learning

Abstract

Drug target interaction is one of the most significant fields of research for drug discovery. The laboratory experiments conducted to identify the drug target interactions are tedious, delayed, and costly. Hence, there is an urgent need to develop highly efficient computational methods for identifying potential drug target interactions that can limit the search space of these laboratory experiments. The existing computational techniques for drug target interaction have been broadly classified into similarity-based methods and feature-based methods. In this paper, a novel feature-based technique to predict drug target interactions has been proposed. The technique uses ensemble learning to determine drug target interactions. Ensemble learning offers greater accuracy in comparison with the traditional classifiers. Thus, the proposed technique aims to improve accuracy using ensemble learning. Also, dimensionality reduction of drug target features is performed using principal component analysis so that the computational time of the method can be reduced. The results indicate an improved performance in comparison with the state-of-the-art methods in the field.