A Study on the Construction of Multiple classifier Systems under the Constrained Number of Classifiers

다국어 초록 (Multilingual Abstract)

Most studies on combining multiple classifiers have mainly focused on how to combine their classification results and only a few studies have investigated on how to construct
multiple classifier systems from the pool of available classifiers. In this paper, an information-theoretic strategy based on information theory model is proposed for constructing the multiple classifier systems. The proposed strategy is applied to the classifiers pool and examines the possible sets of classifiers, and then it selects some sets of classifiers as the multiple classifier system candidates, on the condition that the number of classifiers to be put in the multiple classifier systems is constrained in advance. The multiple classifier system candidates were evaluated and compared together with the constructed sets of classifiers by other strategies in the experiments of the recognition of unconstrained handwritten numerals. The experimental results supported that the proposed strategy was a promising approach.