Summary: Incremental Mixtures of Factor Analysers
Albert Ali Salah, Ethem Alpayd�n
Department of Computer Engineering
Bo�gazic�i University, 34342 Bebek, Istanbul, Turkey
{salah, alpaydin}@boun.edu.tr
Abstract
A mixture of factor analyzer is a semiparametric den�
sity estimator that performs clustering and dimensional�
ity reduction in each cluster (component) simultaneously.
It performs nonlinear dimensionality reduction by model�
ing the density as a mixture of local linear models. The
approach can be used for classification by modeling each
class�conditional density using a mixture model and the
complete data is then a mixture of mixtures. We propose
an incremental mixture of factor analysis algorithm where
the number of components (local models) in the mixture and
the number of factors in each component (local dimension�
ality) are determined adaptively. Our results on different
pattern classification tasks prove the utility of our approach
and indicate that our algorithms find a good trade�off be�

Source: Alpayd�n, Ethem - Department of Computer Engineering, Bogazi�i University

Collections: Computer Technologies and Information Sciences