Title of article :
DESIGNING ROBUST TEMPLATES FOR CELLULAR NEURAL NETWORK BASED ON HYBRID C4.5 ALGORITHM AND GENETIC ALGORITHM
Author/Authors :
Kassem, A. B. Mansoura University - Faculty of Computers and Information Sciences - Department of Computer Sciences, Egypt , Radwan, E. Mansoura University - Faculty of Computers and Information Sciences - Department of Computer Sciences, Egypt , Hamza, T. Mansoura University - Faculty of Computers and Information Sciences - Department of Computer Sciences, Egypt
Abstract :
template design or template learning is a key topic in Cellular Neural Network research, especially design of robust template. Robust tlemplate leads to classifying more objects with high accuracy. This paper presents a novel approach for designing robust templates for both uncoupled and coupled Cellular Neural Network. The learning approach is based on hybrid the machine learning algorithm C4.5 and Genetic algorithm. C4.5 algorithm has been used in order to deduce the optimum structure of template by removing superfluous cells which lead to ieformation redundancy. Genetic algorithm has been used as optimizer for generating parameters of robust template.
Keywords :
Cellular Neural Network , Machine Learning Algorithm C4. 5 , Genetic algorithm , Relative Robustness and Template.
Journal title :
International Journal of Intelligent Computing and Information Sciences
Journal title :
International Journal of Intelligent Computing and Information Sciences
