Role of Machine Learning in Computational Toxicity Prediction

Abstract

It is necessary to do study on how to predict toxicity since actually conducting toxicity testing may be both time-consuming and expensive.
Bioinformatics tools can save time and money. Ever since its start, it has consistently delivered results. The process of analysing and
classifying data is an essential component of bioinformatics. Because of their speed and low cost, in silico approaches have gained
popularity in recent years for evaluating the kinetic and toxic behaviour of drugs. Machine learning is a potent tool for exploring in vitro and
in vivo data for previously undiscovered complicated combinatorial associations. It has found useful applications in areas as varied as
predicting pharmacodynamic characteristics and protein activities, identifying spam, locating oil spills, and recognising human voices.
Algorithms such as Support Vector Machines (SVMs), Artificial Neural Networks (ANNs), and Self Organizing Maps (SOMs), as well as
the difficulties they present, the potential ties they may one day forge, and the web-based toxicity prediction tools have been discussed in
this article.