Using Shannon´s Information Theory and Artificial Neural Networks to Implement Multimode Authentication

Artificial intelligence technologies have been applied in a number of systems to achieve learning and intelligent behavior. In this paper an artificial neural network is used to implement multimode authentication through information fusion. An information fusion model uses metrics computed from the identity attributes using Shannon´s information theory. Initialisation of the artificial neural network is achieved by using the Nguyen-Widrow function while Levenberg-Marquardt back propagation is used as the training algorithm.