Artificial neural network modeling in environmental toxicology

Abstract

Artificial neural networks are increasingly used in environmental toxicology to find complex relationships between the ecotoxicity of xenobiotics and their structure or physicochemical properties. The raison d'être of these nonlinear tools is their ability to derive powerful QSARs for molecules presenting different mechanisms of action. In this chapter, the main QSAR models derived for aquatic and terrestrial species are reviewed. Their characteristics and modeling performances are deeply analyzed. © 2008 Humana Press, a part of Springer Science + Business Media, LLC.