Abstract

The CODES/neural network model has been successfully applied to the prediction of pharmacokinetic properties of therapeutical compounds. The output of CODES, a graphical module based on the Gestalt isomorphism, is proved to be a valuable tool in the design of a neural network model able to predict the half-life values of antihistamines. Additionally, the generated models are able to classify these drugs in their corresponding therapeutic category (H-1 or H-2 receptor antagonists).