International Journal of Soft Computing
ISSN: OnlineISSN: Print 1816-9503
Abstract
In this study, we propose a new fast training algorithm for multilayer perceptron (MLP). This new algorithm is based on the optimisation of Least Fourth (LF) criterion producing a modified form of the Standard Back-Propagation (SBP) algorithm. In this criterion, the Least Fourth error signal is appropriately weighed by the learning coefficient of the steepest descent method. To determine the updating rules for the hidden layers, a similar back propagation method used in the SBP algorithm is developed. This permits the application of the learning procedure to all the neural network layers. Several experiments was carried out indicate significant reduction in the total iteration number, in the convergence time and in the generalization capacities when compared to those of the SBP algorithm.
How to cite this article:
Sabeur Abid and Farhat Fnaiech . Fast Training of Multilayer Perceptrons with Least Mean Fourth (LMF) Algorithm.
DOI: https://doi.org/10.36478/ijscomp.2008.359.367
URL: https://www.makhillpublications.co/view-article/1816-9503/ijscomp.2008.359.367