RTU Research Information System
Latviešu English

Publikācija: Comparative Analisys of Different Approaches Towards Multilayer Percentron Training

Publication Type Publications in RTU scientific journal
Funding for basic activity Unknown
Defending: ,
Publication language English (en)
Title in original language Comparative Analisys of Different Approaches Towards Multilayer Percentron Training
Field of research 2. Engineering and technology
Sub-field of research 2.2 Electrical engineering, Electronic engineering, Information and communication engineering
Authors Aleksandrs Vališevskis
Keywords neural networks; time-series prediction; adaptive learning algorithms; backpropagation; QuickProp; Rprop
Abstract A comparative analysis of four multilayer perceptron learning algorithms is exposed in this work: the error backpropagation algorithm and three other algorithms with fundamentally different approaches towards the improvement of convergence time. Stock exchange share price prediction is at the basis of the comparison of the algorithms. The optimal neural network topology for the solution of the above-mentioned task is determined in this work. Furthermore the forecasts concerning four neural networks with the same topology, but trained with the help of different algorithms are being compared. Special attention is paid to the generalisation ability of neural networks. A series of reasons, which can cause neural network forecast delay problems, is mentioned.
Reference Vališevskis, A. Comparative Analisys of Different Approaches Towards Multilayer Percentron Training. Technologies of Computer Control. Vol.5, 2001, pp.157-167. ISSN 1407-7493.
Full-text Full-text
ID 11124