UCL logo

UCL Discovery

UCL home » Library Services » Electronic resources » UCL Discovery

Flexible and Robust Bayesian Classification by Finite Mixture Models

Archambeau, C; Vrins, F; Verleysen, M; (2004) Flexible and Robust Bayesian Classification by Finite Mixture Models. In: Verleysen, M, (ed.) 12th European Symposium on Artificial Neural Networks, ESANN '2004, Bruges, Belgium, April 28-29-30, 2004. (pp. 75 - 80). d-side: Evere, Belgium.

Full text not available from this repository.

Abstract

The regularized Mahalanobis distance is proposed in the framework of fi nite mixture models to avoid commonly faced numerical difficulties encountered with EM. Its principle is applied to Gaussian and Student-t mixtures, resulting in reliable density estimates, the model complexity being kept low. Besides, the regularized models are robust to various noise types. Finally, it is shown that the quality of the associated Bayesian classifi cation is near optimal on Ripley's synthetic data set.

Type:Proceedings paper
Title:Flexible and Robust Bayesian Classification by Finite Mixture Models
ISBN-13:9782930307046
Publisher version:http://www.dice.ucl.ac.be/Proceedings/esann/esannpdf/es2004-163.pdf
UCL classification:UCL > School of BEAMS > Faculty of Maths and Physical Sciences > Statistical Science

Archive Staff Only: edit this record