KERNEL SELF-ORGANIZING MAP BASED ON RADIAL-BASIS NEURAL NETWORK

  • Евгений Владимирович Бодянский Kharkiv National University of Radio Electronics
  • Анастасия Александровна Дейнеко Kharkiv National University of Radio Electronics
  • Яна Владимировна Куценко Kharkiv National University of Radio Electronics

Abstract

A hybrid artificial neural network is proposed in the present paper. The network combines the principle of kernel systems and self-learning, and is based on radial basis neural networks and self-organizing maps. The proposed system allows to solve the problem of on-line clustering under the conditions, when the classes formed by the initial data have an arbitrary form.

Author Biographies

Евгений Владимирович Бодянский, Kharkiv National University of Radio Electronics

Doctor of Technical Sciences, Professor, Head of PNIL ASU

Анастасия Александровна Дейнеко, Kharkiv National University of Radio Electronics

Junior Researcher

Яна Владимировна Куценко, Kharkiv National University of Radio Electronics

Graduate student

Published
2019-03-02
Issue
No. 20(96) (2015): Electrotechnic and Computer Systems
Section
Computer Systems, Networks and their Components