Multi layer perceptron implementation using matlab

Hello

I am searching how to implement a neural network using multilayer perceptron. My problem is the following : I have a input data matrix with some data for learning and data for test. My intention is to implement the perceptron multilayer algorithm, feed it with these infos and try to tune it sufficiently.

I have some difficulties to understand these points :

- How to enter the matrix data in the neural network ? - What is the stop criteria ? - How can I define precisely the error function to estimate a difference between what is learnt and what was expected ?

Thanks for any help, the best would be an example with lot of explanations.

Best regards.

[Amith Kamath]

would http://www.mathworks.com/help/nnet/examples/index.html#perceptrons or http://www.mathworks.com/products/neural-network/examples.html help?

Answer 1

[Greg Heath]

You can start by using the GUI as explained in the documentation.

That approach can yield a commandline program that can be stored in a file, used, revised and reused.