We would like to obtain an objective measure of how well pairs of clusters
are separated. The clusters are represented by hyperellipsoids with
individual means and covariance matrices. We would like to simulate how
different choices of features affect the normalized distances between the
clusters.
Is there a standard measure of a normalized distance between two
hyperellipsoids, which takes into account the covariant matrices of both
clusters. (Equivalent to the Mahalanobis distance between a point and a
hyperellipsoid?).
Thanks!
Best Regards,
Anders Eikenes
***********************************************************************
Anders Eikenes
Tomra Systems ASA, R&D, Recognition,
Drengsrudhagen 2, P.O. Box 278,
N-1372 Asker, Norway
Email: [log in to unmask]
Phone: +47 66 79 92 91 (direct)
+47 66 79 91 00 (switchboard) Fax: +47 66 79 92 70
***********************************************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|