Skip to Main content Skip to Navigation
Preprints, Working Papers, ...

EM for mixtures - Initialization requires special care

Abstract : Maximum likelihood through the EM algorithm is widely used to estimate the parameters in hidden structure models such as Gaussian mixture models. But the EM algorithm has well-documented drawbacks: its solution could be highly dependent from its initial position and it may fail as a result of degeneracies. We stress the practical dangers of theses limitations and how carefully they should be dealt with. Our main conclusion is that no method enables to address them satisfactory in all situations. But improvements are in-troduced by, first, using a penalized loglikelihood of Gaussian mixture models in a Bayesian regularization perspective and, second, choosing the best among several relevant initialisation strategies. In this perspective, we also propose new recursive initialization strategies which prove helpful. They are compared with standard initialization procedures through numerical experiments and their effects on model selection criteria are analyzed.
Document type :
Preprints, Working Papers, ...
Complete list of metadata

Cited literature [24 references]  Display  Hide  Download
Contributor : Gilles Celeux Connect in order to contact the contributor
Submitted on : Wednesday, February 4, 2015 - 5:09:02 PM
Last modification on : Tuesday, July 13, 2021 - 3:14:08 AM
Long-term archiving on: : Sunday, April 16, 2017 - 8:09:15 AM


Files produced by the author(s)


  • HAL Id : hal-01113242, version 1


Jean-Patrick Baudry, Gilles Celeux. EM for mixtures - Initialization requires special care. 2015. ⟨hal-01113242⟩



Les métriques sont temporairement indisponibles