Abstract: The problems of constructing the selfsrtucturized systems of memory of intelligence information processing tools, allowing formation of associative links in the memory, hierarchical organization and classification, generating concepts in the process of the information input, are discussed. The principles and methods for realization of selfstructurized systems on basis of hierarchic network structures of some special class – growing pyramidal network are studied. The algorithms for building, learning and recognition on basis of such type network structures are proposed. The examples of practical application are demonstrated.

Keywords: knowledge discovery, classification, prediction, growing pyramidal networks, concept formation.

ACM Classification Keywords: I.2.4 Knowledge Representation Formalisms and Methods - Semantic networks, F.1.1 Models of Computation - Self-modifying machines (e.g., neural networks)

Link:

SELFSTRUCTURIZED SYSTEMS

Victor Gladun, Vitalii Velychko, Yurii Ivaskiv

http://www.foibg.com/ijita/vol15/ijita15-1-p01.pdf