Simulating classifier ensembles of fixed diversity for studying plurality voting performance

This paper presents a new method for the artificial generation of classifier outputs in order to analyse the performance of plurality voting according to both the accuracies of the combined classifiers and to the agreement among them. This analysis is conducted in parallel with majority voting in order to compare the efficiency of these two methods when combining dependent classifiers. The experimental results show that the plurality voting is more efficient in achieving the trade-off between rejection rate and recognition rate.