An isometric representation problem in quantum multimolecular polyhedra and similarity: (2) synisometry

Abstract

Collective distances in quantum multimolecular polyhedra (QMP), which can be set as a scalar indices associated to the QMP variance vector, enhance the role of the pair density similarity matrix. This paper describes a simplified efficient algorithm to compute triple, quadruple or higher order density similarity hypermatrices via an approximate isometry: a synisometric decomposition of the pair similarity matrix. Synisometries pretend to avoid the use of Minkowski metrics in QMP description problems, where the double density similarity matrix possesses negative eigenvalues. The synisometric decomposition of the similarity matrix opens the way to the general use of higher order approximate similarity elements in quantum QSAR and in the construction of scalar condensed vector statistical-like indices, for instance skewness and kurtosis. This might lead the way to describe, without excessive complication and within a real field computational framework, the collective structure of quantum multimolecular polyhedra.