Correlative characteristics of the neural network with LMS-algorithm tuning with square constraint with weight vector jitter
The results of statistical analysis of neural network tuning by LMS-algorithm with square constraint for weight vector jitter are represented. The expressions for correlation function of output signal of neuron of the neural of different network layers are obtained. It is shown, that the weight vect...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Saint Petersburg Electrotechnical University "LETI"
2015-04-01
|
| Series: | Известия высших учебных заведений России: Радиоэлектроника |
| Subjects: | |
| Online Access: | https://re.eltech.ru/jour/article/view/18 |
| Summary: | The results of statistical analysis of neural network tuning by LMS-algorithm with square constraint for weight vector jitter are represented. The expressions for correlation function of output signal of neuron of the neural of different network layers are obtained. It is shown, that the weight vector jitter leads to distortions of statistical characteristics of neural network and these distortions increase with increasing of number of neural network layer. |
|---|---|
| ISSN: | 1993-8985 2658-4794 |