A New Feature Selection Method Based On Cla-ec In Face Recognition System
Akhtar Hazrati Bishak, Karim Faez, Morteza Hazrati Bishak · Journal of Mathematics and Computer Science · 2012
Feature selection is an important stage in pattern recognition systems. In this paper we propose a new method based on Cellular Learning Automata- Computing Evolutionary (CLA-EC). The CLA-EC algorithm is an Evolutionary algorithm that is obtained combining from Cellular Learning Automata (CLA) and Computing Evolutionary concept (CA). In this method classification accuracy and number of unselected feature (zeros), considered as fitness function. My Experiments on ORL databases show the effectiveness of the proposed method in compare with Genetic algorithm.