The momentum for training. This is the degree to which changes from which the previous training iteration will affect this training iteration. This can be useful to overcome local minima.
Backpropagation Class | Encog.Neural.Networks.Training.Propagation.Back Namespace