NeuralTools: Training Time for PN/GRN Nets

PN/GRN nets stop when further training won’t improve error; longer training can hurt testing accuracy. Only MLF nets may benefit from longer training.

Jan 22, 2026
Knowledge

Title

Answer

Applies to: NeuralTools, all releases

Can I increase the training time for PN/GRN neural nets?

Answer

No. PN/GRN nets stop training when further error reduction is statistically insignificant. Training longer may lower training-set error but can increase testing-set error (overfitting). Only MLF nets can benefit from longer training, though too many hidden nodes may cause overtraining, so extended training is best used with Best Net Search.

URL Name

training-time-for-pngrn-nets