Comparative Analysis of Classifier Fusers

Abstract

There are many methods of decision making by an ensemble of classifiers. The most popular are methods that have their origin in voting method, where the decision of the common classifier is a combination of individual classifiers’ outputs. This work presents comparative analysis of some classifier fusion methods based on weighted voting of classifiers’ responses and combination of classifiers’ discriminant functions. We discus different methods of producing combined classifiers based on weights. We show that it is not possible to obtain classifier better than an abstract model of committee known as an Oracle if it is based only on weighted voting but models based on discriminant function or classifier using feature values and class numbers could outperform the Oracle as well. Delivered conclusions are confirmed by the results of computer experiments carried out on benchmark and computer generated data.