Two-class classifier cellular automata

Abstract

This paper presents a special class of Cellular Automata (CA) for pattern classification called Two-Class Classifier Generalized Multiple Attractor Cellular Automata (2C2-GMACA). The design is based on two-class classifier architecture using an evolving CA technique to identify a solution. The Generalized Multiple Attractor Cellular Automata (GMACA) is another class of CA for pattern classification. It is better than the Hopfield Net in literature. In addition, it is compared with the 2C2-GMACA in performance evaluation. According to the Error Correcting Codes experiment, the 2C2-GMACA is more powerful than the GMACA in term of recognition rates and evaluation time to get a rule vector which is reduced to linear complexity.