Volume 2, Number 3/4, p. 80-84
N. Tkemaladze
Department of Classification of Complex Systems, Institute of Cybernetics, 5 Sandro Euli St, 380086 Tbilisi, Georgia
On the problems of an automated system of pattern recognition with learning
An automated system of pattern recognition with learning (PRL) is elaborated. In this article the theoretical principles of the PRL system are given. The functioning of this system does not depend on the physical essence of the objects (patterns) under investigation, but merely on the type and number of initial parameters. The PRL system consists of two subsystems, the learning model (LM) and the model of recognition (MR). The PRL system was successfully used for the recognition of objects of different classes. In many cases the system is able to solve complicated problems of pattern recognition on the basis of minimal initial information.
Keywords: knowledge and data bases, learning model, PRL