Robust QSAR models using Bayesian regularized neural networks

We describe the use of Bayesian regularized artificial neural networks (BRA NNs) in the development of QSAR models. These networks have the potential t o solve a number of problems which arise in QSAR modeling such as: choice o f model; robustness of model; choice of validation set; size of validation effort; and optimization of network architecture. The application of the me thods to QSAR of compounds active at the benzoaiazepine and muscarinic rece ptors is illustrated.