Kohonen’s Self Organizing Map:

• Kohonen self organizing maps refers also known as Kohonen feature maps or topology preserving maps are another competition based network pradigm for data clustering.
• Networks of this type impose a neighbourhood constraint on the output units, such that certain topological property in the input data is reflected in the output units weights
• The learning procedure of Kohonen feature maps is similar to that of competitive learning networks i.e. a similarity (dissimilarity) measure is selected and the winning unit is considered to be the one with the largest (smallest) activation.
• For Kohonen feature maps, we update not only the winning unit’s weights but also of the weights in a neighbourhood around the winning units.
• If the output units of a competitive learning network are arranged in a geometric manner (such as a 1D vector or 2D array) then we can update the weights of the winners as well as the neighbouring losers.