he classifiers characterized by a tagged feature-class representation, a univariate discrimination approach, a cooperative classification scheme, and a logic-based learning strategy are discussed. Neither of the classifiers bears the constraints to the fixed sets of features and classes. Concepts of the tagged feature-class representation and the properties of feature matching in the dynamic environment are studied. Experimental tests and results of the classifiers are illustrated.
Available at: http://works.bepress.com/qiuming-zhu/33/