R help - Apr 2010 - Optimize nnet() for cross-validation error

I would run 200 iterations of nnet() such that optimization is done for
'f',
where f = error_cross_validation * (error_cross_validation /
error_training_set). Any guidance on how to optimize for this per-cycle
dependent variable (f) is appreciated.

Thank you,
supra

	[[alternative HTML version deleted]]