Improving classification accuracy using cellular automata

Abstract

Ensembles of classifiers have the ability to boost classification accuracy comparing to single classifiers and are a commonly used method in the field of machine learning. However in some cases ensemble construction algorithms do not improve the classification accuracy. Mostly ensembles are constructed using specific machine learning method or a combination of methods, the drawback being that the combination of methods or selection of the appropriate method for a specific problem must be made by the user. To overcome this problem we invented a novel approach where ensemble of classifiers is constructed by a self-organizing system applying cellular automata (CA). First results are promising and show that in the iterative process of combining the classifiers in the CA, a combination of methods can occur, that leads to superior accuracy. © Springer-Verlag 2004.