Beginning a new neural network from previously trained weight

I'm pretty new to matlab and the neural network toolbox, so I'm not sure how to play around with the right commands, but I need to avoid overfitting my data. It was proposed to me that I run a neural network trial with the training input, training output and the data for which I would like a predicted output. I think what I need to do first is run a neural network first, change my input/output data and start a new neural network BEGINNING with the weights of the previous trained neural network. That way I can use a wide array of training data to converge to a more accurate prediction. I would like the cod I would need to do this task in matlab, having already written the neural network script, considering that I have an inputfile,outputfile, and file for which the output is to be predicted.

Answer 1

[Greg Heath]

Sounds like a waste of time.

 1. See the help and doc examples

 2. Search for posted examples in NEWSGROUP and ANSWERS

 greg networkname Ntrials

For example

 greg fitnet Ntrials

Thank you for formally accepting my answer

Greg