Help with plotting Neural Network training / testing results

I'm trying to plot the results of the training of my neural network and the test results when i simulate the network with test data.

I am plotting using:

plot(ptr, ttr, '-', ptr, an, '-.')
legend('Target', 'Network output')
title('Approximation performance with training data');

But I am getting the error:

Error using plot
Vectors must be the same lengths.
Error in task2V2 (line 264)
plot(ptr, ttr, '-', ptr, an, '-.')

ptr is 63x10Double ttr is 10x10Double and an is 10x10Double

ptr = training data ttr = the target output an = simulating the trained network with training data

any ideas or pointers would be helpful

Answer 1

[Greg Heath]

[I N ] = size(p)

[ O N ] = size(t)

What are your values???

Are you using fitnet for regression or patternet for classification?

Assuming regression with all of the fitnet defaults(including the trn/val/tst default ratio of 0.7/0.15/0.15)

net = fitnet; [ net tr y e ] = train(net, p, t);

All of the training, validation and test results are summarized in the output on the LHS:

tr = tr

See the documentation on how to retrieve them.

Since inputs and outputs are multidimensional, you will have 6*I*O plots which, obviously, is not practical. (Each output vs input pairing yields plots of target and output vs input for trn, val and test)

You probably do not want to do this unless I and O dimensionalities are small.

Click on the training tool window to see the default plots.

Hope this helps.

Thank you for formally accepting my answer.

Greg