Hybrid algorithm with variants for feedforward neural network
Joint Authors
Kandasamy, Thinakaran
Rajendran, Rajasekar
Source
The International Arab Journal of Information Technology
Issue
Vol. 15, Issue 2 (31 Mar. 2018)6 p.
Publisher
Publication Date
2018-03-31
Country of Publication
Jordan
No. of Pages
6
Main Subjects
Information Technology and Computer Science
Abstract EN
Levenberg-Marquardt back-propagation algorithm, as a Feed forward Neural Network (FNN) training method, has some limitations associated with over fitting and local optimum problems.
Also Levenberg-Marquardt back-propagation algorithm is opted only for small network.
This research uses hybrid evolutionary algorithm based on PSO in FNN training.
This algorithm includes a number of components that gives advantage in the experimental study.
Variants such as size of the swarm, acceleration coefficients, coefficient constriction factor and velocity of the swarm are proposed to improve convergence speed as well as to improve accuracy.
The integration of components in different ways in hybrid algorithm produces effective optimization of back propagation algorithm.
Also, this hybrid evolutionary algorithm based on PSO can be used for complex neural network structure
American Psychological Association (APA)
Kandasamy, Thinakaran& Rajendran, Rajasekar. 2018. Hybrid algorithm with variants for feedforward neural network. The International Arab Journal of Information Technology،Vol. 15, no. 2.
https://search.emarefa.net/detail/BIM-838606
Modern Language Association (MLA)
Kandasamy, Thinakaran& Rajendran, Rajasekar. Hybrid algorithm with variants for feedforward neural network. The International Arab Journal of Information Technology Vol. 15, no. 2 (Mar. 2018).
https://search.emarefa.net/detail/BIM-838606
American Medical Association (AMA)
Kandasamy, Thinakaran& Rajendran, Rajasekar. Hybrid algorithm with variants for feedforward neural network. The International Arab Journal of Information Technology. 2018. Vol. 15, no. 2.
https://search.emarefa.net/detail/BIM-838606
Data Type
Journal Articles
Language
English
Notes
Includes appendix.
Record ID
BIM-838606