Unified decision combination framework

Pattern recognition researchers are discovering that a judicious combination of classifiers, generally outperforms a single one. In this paper, we present a unified framework for decision combination. By looking at the combination problem from this new perspective, we gain an intuitive, perhaps better understanding of some of the known algorithms. A new parameterized combination method (pooled ranking figure of merit) is presented which is shown to be equivalent to three of the standard combination methods. Furthermore, our new decision combination method provides newer combination methods that provide interesting tradeoffs. These ideas are supported by simulation experiments. (C) 1998 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.