Hi, I'm Zaidi and doing my Phd in Intelligent System at Cardiff University, UK. I have tested a control chart pattern recognition data with MLP, RBF and LVQ. Next I want to test with SVM. My question here is what is the different between SVM and RBF/MLP/LVQ network? How to relate the support vector equation with the SVM network diagram. Start from input layer, between input and hidden layer, in hidden layer and between hidden to output layer. I've got my own training algorithm to test with svm. How do we know the suitable number of hidden layer? Thanks for help Zaidi Cardiff University, UK