ROMJIST Volume 23, No. T, 2020, pp. T78-T93
Valerii ANIKIN, Oksana ANIKINA, Oksana GUSHCHINA Kohonen Neural Network Training by Cellular Automaton
ABSTRACT: Effective integral dispersion quantile measure of the quality of data topology mapping by the Kohonen neural network is proposed. The comparison of the algorithm of SOM training by cellular automaton and classical Kohonen online and batch training algorithms was systematically carried out. Advantages and disadvantages of the new SOM training algorithm were identified.KEYWORDS: Artificial Intelligence, Kohonen neural network, self-organizing map, learning algorithm of SOM, cellular automaton, quantile dispersive measureRead full text (pdf)
