Hierarchical pattern classification (Q1113691)

The cluster validity problem, using a hierarchical classification method is presented and an algorithm based on this method is given. For the set of the patterns X no a priori knowledge about its cluster structure is supposed to be given. X is decomposed into the fuzzy partition classes. The classes are subdivided as long as it is necessary to produce the final objective classification due to the minimization criterion. Every class (cluster) is presented by its center. Application for feature classification process is proposed.