Classic SVM

The classic (linear) SVM is similar in use and inference to the classic perceptron, and is a special case of an Artificial Neuron (AN) method. Inference involves applying a transfer function to an input vector, and feeding it into an activation function, just like any AN. Classic SVMs are all but guaranteed to have maximum performance for binary, linearly separable problems. What separates an SVM from the other AN methods is the way the activation function is defined, and correspondingly what weights will be learned.