What is an RBF?

Radial Basis Function is a type of Multi Layer Perceptron which has one input layer, one output layer, and one hidden layer: Here we use a radial basis function as the Activation Function, we commonly use the Gaussian function.

Why use RBF?

To convert non-linearly separable data into linearly-separable, or function interpolation (i.e. predict function) non-linearly.

How to Train?

You can either initialize nice clusters and then train the weights, or apply a clustering algorithm first then train weights.