Title :
Optimal dead-zone characteristic for minimizing the a-posterior error in basis function networks
Author_Institution :
Tech. Univ. Wien, Austria
Abstract :
The incorporation of dead-zones in the error signal of basis function networks avoids the networks´ over-training and guarantees the convergence of the normalized LMS-algorithm and related algorithms. A new so-called error-minimizing dead-zone is presented providing the least a-posteriori error out of the set of all convergence assuring dead-zones
Keywords :
convergence of numerical methods; error analysis; feedforward neural nets; least mean squares methods; minimisation; a-posteriori error minimization; basis function networks; convergence-assuring dead-zones; error-minimizing dead-zone; least a-posteriori error; normalized LMS-algorithm convergence; optimal dead-zone characteristic; Contracts; Convergence; Electronic design automation and methodology; Error correction; Europe; Intelligent networks; Iterative algorithms; Linear approximation;
Conference_Titel :
Decision and Control, 1996., Proceedings of the 35th IEEE Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-3590-2
DOI :
10.1109/CDC.1996.574357
