Chapter 5 Similarity/Diversity Measure for Sequential Data Based on Hasse Matrices: Theory and Applications

Publisher Summary This chapter describes a new similarity/diversity measure as a new approach for the analysis of sequential data, where useful information can be obtained by the ordering relationships between the sequence elements. This new similarity/diversity measure is based on the distance evaluated between pairs of Hasse matrices derived from the classical partial ordering rules. It can be naturally standardized, thus allowing the interpretation of these distances as absolute values and deriving simple similarity and correlation indices. The similarity/diversity between two sequences is obtained by the definition of a distance between the corresponding Hasse matrices. These distances have some useful properties and seem to show high sensitivity to changes in structure sequences. This methodology can be used for several applications: (a) evaluation of molecular similarity/diversity, using sets of sequential descriptors, (b) evaluation of similarity between spectra or sequential analytical data, and (c) evaluation of DNA and protein sequences and, in general, assessment of similarity of sequential data.