The following information was submitted:
Transactions: WSEAS TRANSACTIONS ON SIGNAL PROCESSING
Transactions ID Number: 32-707
Full Name: Mohamed salah Salhi
Position: Ph.D. Candidate
Age: ON
Sex: Male
Address: 3 rue 6988 El omrane sup. tunis
Country: TUNISIA
Tel: 98469287
Tel prefix:
Fax:
E-mail address: salhi_mohamedsalah@yahoo.fr
Other E-mails:
Title of the Paper: A suitable model of evolutionary SOM for phonemes recognition
Authors as they appear in the Paper: Mohamed Salah Salhi, Najet Arous, Noureddine Ellouze
Email addresses of all the authors: salhi_mohamedsalah@yahoo.fr;Najet arous@yahoo.fr;N. Ellouze@enit.rnu.tn
Number of paper pages: 12
Abstract: The neural networks, particularly Kohonen maps known by SOM, are considered as nonlinear parametric models for universal recognition which is highly qualified in the analysis of great volumes data flow. But what is undeniable that these models offer a limited recognition rates to the local optima. In another case, the application of genetic algorithm GA, that's in fact inspired from natural selection and genetics, has a premature convergence, but it may sometimes have inexact results. These problems can be solved through a promising idea which consists in introducing the approaches, making the hybridization of the two models SOM and GA to get profit from their common advantages. This paper, suggests a general representation of various approaches that are the target of this hybridization. It's highly important to bear in mind that we are introducing a method which guaranties population diversity over generations. By intervening on the possible SOM parameters, t!
he genetic algorithms brought to this hybridization the property of total research. The obtained architectures make it possible to simulate the human intelligence in a human way.
Keywords: Kohonen Map SOM, Genetic Algorithms GAs, hybrid Models GA-SOM
EXTENSION of the file: .pdf
Special (Invited) Session:
Organizer of the Session:
How Did you learn about congress:
IP ADDRESS: 196.203.130.3
