Methods for updating and training for a self-organising card

A method for updating a self-organising map comprising neurons and valued connections between the neurons, each of the neurons being associated with a respective vector, referred to as the prototype vector, the method comprising:

• selecting the two neurons (u* and u**), referred to as the best winning neuron and second best winning neuron, respectively, each having the prototype vector, respectively, the first and second closest to a vector of a learning point (x(k)), in accordance with a determined distance rule,
• modifying the prototype vectors of the best winning neuron and the neurons located around the best winning neuron in the self-organising map, in the direction of the vector of the learning point (x(k)),
• determining the neighbouring neurons (N(u*)) of the best winning neuron (u*) in the self-organising map,
• if the second best winning neuron (u**) is part of the neighbouring neurons (N(u*)) increasing the valuation of the connection between the first and second best winning neurons, the method being characterised in that it comprises:
• reducing the valuation of each connection between the first best winning neuron and the direct neighbouring neurons (N(u*)) other than the second best winning neuron (u**), and
• wherein the method steps are performed by a processor.