A modified learning algorithm incorporating additional functional constraints into neural networks
Publication in refereed journal

Times Cited
Web of Science2WOS source URL (as at 21/10/2020) Click here for the latest count
Altmetrics Information

Other information
AbstractIn this paper, a modified learning algorithm to obtain better generalization performance is proposed. The cost terms of this new algorithm are selected based on the second-order derivatives of the neural activation at the hidden layers and the first-order derivatives of the neural activation at the output layer. It can be guaranteed that in the course of training, the additional cost terms for this algorithm can penalize both the input-to-output mapping sensitivity and the high frequency components to obtain better generalization performance. Finally, theoretical justifications and simulation results are given to verify the efficiency and effectiveness of the proposed learning algorithm.
All Author(s) ListHan F, Li XQ, Lyu MR, Lok TM
Name of ConferenceInternational Conference on Intelligent Computing
Start Date of Conference23/08/2005
End Date of Conference26/08/2005
Place of ConferenceHefei
Journal nameInternational Journal of Pattern Recognition and Artificial Intelligence
Detailed descriptionTo ORKTS: indexed by SCI
Volume Number20
Issue Number2
Pages129 - 142
LanguagesEnglish-United Kingdom
Keywordsconstrained learning algorithm; feedforward neural networks; generalization performace; time series prediction
Web of Science Subject CategoriesComputer Science; Computer Science, Artificial Intelligence; COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Last updated on 2020-22-10 at 01:15