A Machine Learning-Based Method for Predicting unknown Pharmacointeractions

Abstract

A lot of research has been done on the efficacy of machine learning algorithms in predicting the pharmacological interference between two drugs. Ordinarily, this interference depends on many factors such as the taxonomical, chemical, pharmacological or genomic similarities between the two drugs. Nevertheless, a lot of adverse events (AEs) are reported every year, due to the simultaneous consumption of two or more drugs. Much research has been conducted on the accuracy of the interference prediction based on these factors, each differing in the algorithms and factors used. In this publication, we propose a machine learning-based approach to predict undiscovered drug-drug interactions based on a few of the impacting factors, for better results and thus, help minimize the potential harm that can be caused to society.