Title :
A cloning approach to classifier training
Author :
Al-Alaoui, Mohamad Adnan ; Mouci, Rodolphe ; Mansour, Mohammad M. ; Ferzli, Rony
Author_Institution :
Dept. of Electr. & Comput. Eng., American Univ. of Beirut, Lebanon
fDate :
11/1/2002 12:00:00 AM
Abstract :
The Al-Alaoui algorithm is a weighted mean-square error (MSE) approach to pattern recognition. It employs cloning of the erroneously classified samples to increase the population of their corresponding classes. The algorithm was originally developed for linear classifiers. In this paper, the algorithm is extended to multilayer neural networks which may be used as nonlinear classifiers. It is also shown that the application of the Al-Alaoui algorithm to multilayer neural networks speeds up the convergence of the back-propagation algorithm.
Keywords :
backpropagation; character recognition; mean square error methods; neural nets; pattern classification; Al-Alaoui algorithm; Bayes classifier; Levenberg-Marquardt algorithm; character recognition; mean-square error; multilayer neural networks; neural networks; pattern classification; pattern recognition; Algorithm design and analysis; Approximation algorithms; Character recognition; Classification algorithms; Cloning; Convergence; Multi-layer neural network; Neural networks; Pattern classification; Pattern recognition;
Journal_Title :
Systems, Man and Cybernetics, Part A: Systems and Humans, IEEE Transactions on
DOI :
10.1109/TSMCA.2002.807035
