Converting target values into form appropriate for Neural Network Classificaton

I have a dataset in which class attribute has string values like normal,neptune,smurf etc against every tuple. I want to train a neural network to classify inputs into one of these target values. I want to convert these string class values into form like [0 0 1] [0 1 0] [1 0 0] means if the class is smurf then its [0 0 1] if its neptune then its [1 0 0]. Can someone tell how to do this in matlab.

Answer 1

[Greg Heath]

1. See the patternnet help and documentation examples

 help patternnet
 doc patternnet

Search the NEWSGROUP and ANSWERS for my patternnet posts

 greg patternnet

Hope this helps.

Thank you for formally accepting my answer

Greg

[Garry Maker]

Thanks Greg. I converted my target values using ind2vec function. Now what happen is when ever I tries to train my network (patternnet) Matlab throws me an error " Matlab encountered an internal error please close and restart" something like this. Even though I generalized classes to 5 from 23 but still same error. Please help.

[Greg Heath]

What happens when you use one of the classification dataset examples in the documentation?

 help nndatasets

 doc nndatasets

[Garry Maker]

ind2vec returned sparse double matrix so I have to convert it into full using full(s) and it worked. Thanks for your concern.