Moment matching estimation method for an asymmetric generalized Gaussian mixture model

Auteurs :  N. Nacereddine, D. Ziou
Année : 2013
Domaine : Electronique
Type : Communication
Conférence: 11th Intern. Conf. Pattern Recognition and Image Analysis: New information technologies, Samara, Russia.
Résumé en PDF :  (résumé en pdf)
Fulltext en PDF :  (.pdf)
Mots clés : Asymmetric generalized Gaussian distribution, finite mixture model, moment matching method, EM algorithm

Résumé : 

In this paper, the r-order moments of the asymmetric generalized Gaussian (AGG) distribution is originally computed. Then, the moment matching method associated to the expectation-maximization (EM) algorithm is used to estimate the AGG mixture model parameters. The obtained results are comparable to those of the maximum likelihood method which, however, manipulates high nonlinear equations (piece-wise function, log, etc.), contrarily to the proposed method where the calculus is less difficult.