Can I use neural network toolbox for Probabilistic Neural Network? If not, how do I run it with code?

Hi,

I've got CSV data that lines up with 12 data columns:

ID A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12

W1 360 331 131 229 361 10.0368 2.855 251.51 0 0.377049 1.526492 2

I've tried putting column one as text or as numeric and numeric works better.

What I want to do is run a probabilistic neural network so that I can have different classification outcomes. The first column is identity, the next 12 are variables describing identity. The individuals in the validation database may OR may NOT be present in the training database and I need it to classify to either "known identity = W1, W2" etc or "unknown identity". I know this works with MLP Artificial Neural Networks in Statistica so that the information should be enough to classify them, but Statistica doesn't do PNNs.

I'm a novice Matlab user though and I'm not sure how to use it for this. I've tried using the Neural Network toolbox but it doesn't let me load targets for the classification option. And I can't see where the PNN option would come.

I've tried using the newpnn code and get stuck on how to execute.

Sorry this is so basic but I don't have anyone at my uni who has done this and can help.

Thanks in advance,

Holly

[Greg Heath]

It would be much easier if you could post your WELL FORMATTED code with comments and error messages.

[Greg Heath]

Especially clarify what this means:

 "but it doesn't let me load targets for the classification option"