Design of radial basis probabilistic neural network

I want to design a RBPNN for pattern recognition using the recursive orthogonal least squares for training weights output and selecting hidden centers in the first hidden layer. I don't know to write this code. I have in input layer two neurons that represent the coordinates of the points, and I have five classes that I want to classify.Help me! Thanks!

Answer 1

[Greg Heath]

I'm not familiar with that method. The only MATLAB option is NEWRB. If you want to use it, see my posts in MATLAB NEWSREADER.

Hope this helps.

Greg

[Greg Heath]

P.S. The input nodes are not neurons. Only hidden and output layer nodes are neurons.

[Greg Heath]

Also, remove the tag rbpnn.