A Differential Evolution-based Rule Ordering of Cellular Automata for Classification

This paper proposed a differential evolution-based rule ordering of cellular automata for classification, called DEROCA. DEROCA aims to improve the performance of Cellular Automata-based Classmer (CAC) previously presented. CAC faced a problem of rule ordering in classification. It utilized Genetic Algorithm (GA) which reported the drawback when copes with complicated problems with high dimension space. This paper gets rid of such a limitation by computing Differential Evolution (DE) in the phase of rule ordering. For performance evaluation, a number of datasets consisting of binary and non- binary features was empirically experimented in comparison with CAC and Support Vector Machines (SVM). In this regard, DEROCA provided the outstanding results reporting the outstanding performance on average better than CAC and SVM, respectively.