Classification of binary vectors by using ΔSC-distance

Stochastic complexity (SC) has been employed as a cost function for solving binary clustering problem. Shannon code length (CL-distance) has been previously applied for the purpose of classifying the data vectors during the clustering process. The CL-distance, however, is defined for a given (static) clustering only, and it does not take into account of the changes in the class distribution during the clustering process. We propose a new ΔSC-distance function based on a design paradigm, in which the distance function is derived directly from the difference of the cost function value be re and after the classification. The ΔSC is general in the sense that it does not depend on the algorithm in which it is applied The effect of the new distance function is demonstrated by implementing it with the GLA and the RLS clustering algorithms.