Skoči na glavni sadržaj

Izvorni znanstveni članak

https://doi.org/10.19279/TVZ.PD.2020-8-2-07

THE IMPACT OF NOISE ON THE LEARNING OF A SINGLE NEURON

Predrag Valožić orcid id orcid.org/0000-0001-7177-0692 ; Tehničko Veleučilište u Zagrebu, Zagreb, Hrvatska


Puni tekst: hrvatski pdf 3.603 Kb

str. 114-120

preuzimanja: 192

citiraj


Sažetak

The behaviour of a linear neuron learning by template contaminated by additive white Gaussian noise was analysed. All calculations and graphs were performed in Matlab.The digital filtering of prediction errors in the Widrow - Hoff learning algorithm achieves reliable convergence of the learning process. The final results of the sine wave generator design are of better quality.The conclusions are also implicit in daily, human situations.

Ključne riječi

Linear neuron; Noise; Learning rule; ADALINE; Sine-wave generator

Hrčak ID:

243891

URI

https://hrcak.srce.hr/243891

Datum izdavanja:

15.9.2020.

Podaci na drugim jezicima: hrvatski

Posjeta: 1.110 *