NeuroCOLT workshop
on
Applications of Learning to Text and Images
Windsor, 30 April - 2 May 2001
Cumberland Lodge

On Classification Using Radial Basis Function Neural Networks

Michael Schmitt

mschmitt@lmi.ruhr-uni-bochum.de

A new method for classifying data with radial basis function (RBF) neural networks is presented that uses the capabilities of RBF units more effectively. The method arose from considerations of biological neuron models. It is suitable for high-dimensional, real-valued data and can be incorporated into many existing learning algorithms for RBF neural networks.