Experimental Comparison Between Implicit and Explicit MCSs Construction Methods

Multiple classifier machines (MCSs) is a very popular research topic in recent years. It has been proved theoretically and empirically to outperform single classifiers in many scenarios. Creating diverse sets of classifier is one of the key issues in MCSs. One kind of method measures the diversity among the individual classifier when building the MCS while the other method does not consider the diversity value directly. This paper compared these two kinds of methods experimentally. From the experiments, the performances of implicit and explicit methods are very close. We can conclude that it is not necessary to consider the diversity measure among individual classifiers directly for building a good MCS