How is test set used in trainbr algorithm?

I am using trainbr as the training algorithm for my neural network. What I am trying to determine is if the test set is being used in the network optimization process, or is it being left alone until the end? For instance, the code is simply:

net = newff(inputs,targets,n);

net = init(net);

[net tr] = trainbr(net,inputs,targets);

Is the test set being used during each iteration in some validation sense, or is separate until the end. I ask because I have used this technique and also trained only the first 75% of my data using the net.divideFcn = ''. When I do this and then run the test set with the network, I get a different response. I just wanted to make sure that the general trainbr function does not use the test set during the development of the network (i.e. weights, etc.).

Thank you very much for your help.

Answer 1

[Greg Heath]

The test set is never used to influence the design.

Thank you for formally accepting my answer

Greg

[Nirvana Delgado]

Dear Greg,

thank you for your answer. Would you mind elaborating on that?

What is the test set used for when using trainbr?

Thank you in advance!