Model-based mixture discriminant analysis an experimental study
Zohar Halbe and Mayer Aladjem
Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev
P.O.Box 653, Beer-Sheva, 84105, Israel
The subject of this paper is an experimental study of a discriminant analysis (DA) based on Gaussian mixture estimation
of the class-conditional densities. Five parameterizations of the covariance matrixes of the Gaussian components are
studied. Recommendation for selection of the suitable parameterization of the covariance matrixes is given.
Keywords: Discriminant analysis, Gaussian mixture model, Density estimation, Model selection.
Discriminant analysis (DA) is a powerful technique for classifying observations into known pre-
existing classes. In the Bayesian decision framework  a common assumption is that the observed d-
dimensional patterns x (xRd
) are characterized by the class conditional density fc(x), for each class
c=1, 2,..., C. Let Pc denotes a prior probability of the class c. According to Bayes theorem the posterior
probability that an arbitrary observation x belongs to class c is
( | ) ( )/ ( ).