A Survey on computational methods for the classification of compounds into drug and non-drug

Abstract

The introduction of computer methods and machine learning has drastically changed the field of drug discovery. Now, drug
designing has evolved from Drug Discovery into Computer Aided Drug Discovery (CADD). There are several reasons that
led to this change. The low success rate of classification of compounds into drugs and non-drugs, being one of those
reasons, has evoked a renewed interest in understanding more clearly what makes a compound drug-like and its new model
and methods for classification. Our work reviews a number of works based on different methods and models of machine
learning techniques for identifying and classifying drug-like molecules which ranges frombasicdescriptorbasedmethodtothelatestfingerprint-basedmethod.