Similar Items: Comparison of Bayesian learning and conjugate gradient descent training of neural networks