Dependency Relationship Based Decision Combination in Multiple Classifier Systems

Although many decision combination methods have been proposed, most of them did not focus on dependency relationship among classiiers in combining multiple decisions. That makes classiication performance of combining multiple decisions be degraded and biased, in case of adding highly dependent inferior classiiers. To overcome such weaknesses and obtain robust classiication performance, the present study used dependency relationship for better combining multiple decisions. In order to identify dependency relationship by observing outputs of multiple classiiers, two methods are used on the basis of rst-order dependency relationship. One is to use the concept of mutual information , and the other one is to use the concept of statistically measured association. The rst-order dependencies identiied are used to combine multiple decisions, using Bayesian formalism. A number of multiple classiier systems are applied to totally uncontrained on-line handwritten numerals and the English alphabet recognition. The experimental results show that the classiication performance of a multiple classiier system is superior to that of individual classiiers. Also, they show that considering the dependency relationship outperforms others in accuracy, when the highly dependent inferior classiiers are added.