how to set target for neural network

I had extracted feature vector of a signal and saved it in a excel document. feature vector is 128*1922 dimension. I don't know how to create target for this input so i can train the neural network. I need to do signal classification. 35 classes. So, how do i create target vector and train the network?

Answer 1

[Greg Heath]

 [ I N ] = size(input) % [ 128 1922 ]
 [ O N ] = size(target) %[  35 1922 ]

Where target columns are columns of the unit matrix eye(35). The relationship between trueclassindices in {1:35} and the corresponding target is

 target = full(ind2vec(trueclassindices))
 trueclassindices = vec2ind(target)

Estimated output indices are obtained from

 output = net (input);
 estimatedindices = vec2ind(output);

with

 errors = estimatedindices ~= trueclassindices;
 Nerrs = sum(errors)
 Pcterr = 100*Nerrs/N

I have posted MANY detailed examples in BOTH the NEWSGRIOUP and ANSWERS.

Hope this helps.

Thank you for formally accepting my answer

Greg

Answer 2

[Walter Roberson]

Create a vector of length 1922 (I assume that is the number of samples you have), in which each value is a member of 1:35 depending on which of the classes the sample corresponds to.

For some of the NN routines, you need to use a binary array rather than a class number. Which NN routine are you using?