What is validation failure Artificial neural network ?

while using nntraintool the training stops when no of validation failures reach certain prescribed number. What exactly is this Validation failure? Also, what is the term like Mu stands for?

Answer 1

[Greg Heath]

To prevent the net from performing poorly on nontraining (validation, test and unseen data) while learning well on training data, training stops if the validation performance degrades for 6 (default) consecutive epochs. The procedure is called

 VALIDATION STOPPING.

Hope this helps.

Thank you for formally accepting my answer

Greg

[Akhil Reddy]

What is Mu?

[Greg Heath]

Mu is the momentum constant. To see how it's used lookup the explanation of the Levenberg-Marquardt algorithm.