Self-organizing map and its learning in the fuzzy clusteringclassification tasks

Abstract

In the paper, combined self-learning and learning method of self-organizing map (SOMLVQ) is proposed. Such method allows to increase quality of information processing under condition of overlapping classes due to rational choice of learning rate parameter and introducing special procedure of fuzzy reasoning in the clustering-classification process, which occurs both with external learning signal (“supervised”), and without one (“unsupervised”). As similarity measure of neighborhood function or membership one, cosine structures are used, which allow to provide a high flexibility due to self-learning-learning process and to provide some new useful properties.