OPTIMIZING THE CLUSTERING PERFORMANCE OF A SELF-ORGANIZING LOGIC NEURAL-NETWORK WITH TOPOLOGY-PRESERVING CAPABILITIES

In this article, a self-organising logic neural network is studied. Th is network successfully clusters input patterns into classes character ised by a high similarity, while assigning these classes to the networ k nodes so that relationships existing in the pattern space are replic ated on the network structure. The network performance is optimised by (i) introducing a mechanism which ensures the efficient use of the ne twork nodes for storage of pattern classes and by (ii) determining the training strategy which results in optimal topology-preservation char acteristics.