Cluster analysis is the production of a classification scheme for initially unclassified and
unlabelled data; the aim being that the classification produced should represent the
natural groups/clusters that may be present in the data.
Unsupervised clustering using neural networks is a well established technique, with the
simple competitive learning neural network (CLNN) being the elementary model and
Kohonen’s self organising maps (SOMs) the most popular manifestation [Hertz et al.,
1991].