Things you haven't heard about the self-organizing map

The self-organizing map (SOM) algorithm can be related to a biological neural network in many essential known details; even cyclic behavior automatically ensues from a simple nonlinear neural model, whereby these cycles correspond to the steps of the discrete-time SOM algorithm. Compared with the other traditional neural-network algorithms, the SOM alone has the advantage of tolerating very low accuracy in the representation of its signals and synaptic weights. This is proven by simulations. Such a property ought to be shared by any realistic neural-network model. While the SOM can thus be advanced as a genuine neural-network paradigm, it is shown how the basic algorithm can be generalized and made more computationally efficient in several ways.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>