New Meta Classifier Beats Competition, Boosts Accuracy in Pattern Recognition
A new classifier called Rough Ensemble Classifier (RSM) was developed by combining multiple classification systems. The RSM was compared to the Bayes classifier, which is considered the best. Results showed that the RSM performs as well as the best individual classifier and even converges to Bayes if one of its base classifiers does. Experimental tests demonstrated that the RSM significantly improves classification accuracy compared to other ensemble methods.