Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Properties of Convergence of Weights of the Autoassociative Two-Layer Linear Neural Network at Construction of Compressing Displays of Casual Vectors

Keywords:

A.A. Sirota, V.G. Popov


The theorem and consequence to it of convergence of weight factors of an autoassociative linear neural network with reduced number of neurons in the latent layer to basic functions of Karunen-Loev decomposition is proved at training on set of realizations of a casual vector. Properties of convergence are investigated at training a network on realizations of casual vectors with the set correlation characteristics
References:

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio