A New Symbolic Dissimilarity Measure for Multivalued Data Type and Novel Dissimilarity Approximation Techniques

In this paper a new statistical measure for estimating the degree of dissimilarity between two symbolic objects whose features are multivalued symbolic data type is proposed. In addition two new simple representation techniques viz., interval type and magnitude type for the computed dissimilarity between the symbolic objects are introduced. The dissimilarity matrices obtained are not necessarily symmetric. Hence, clustering algorithms to work on such unconventional approximated matrices, by introducing the concept of mutual average dissimilarity value and magnitude average dissimilarity respectively for interval type and magnitude type approximation representations are also proposed.