[R] em algorithm mixture of multivariate normals

Christian Hennig chrish at stats.ucl.ac.uk
Thu May 21 12:41:04 CEST 2009

Look up packages flexmix and mclust!


On Thu, 21 May 2009, daniele riggi wrote:

> Hi,
> I would like to know if it is possible to have a "R code" to estimate the
> parameters of a mixture of     bivariate (or multivariate) normals via EM
> Algorithm. I tried to write it, but in the estimation of the matrix of
> variance and covariance, i have some problems. I generate two bidimensional
> vectors both from different distribution with their own vector means and
> variance and covariance structure. When I create a unique vector, the
> structure of covariance changes, and so the implementation of the EM
> algorithm doesn't work.
> Maybe someone knows the reason. If I fix the starting initial value of the
> covariance matrix and I don't update the estimate of this matrix, the
> algorithm works and finds the estimate of the vector means, so I wrote it in
> the correct way. However if someone could help me I will be very grateful to
> him for kindness.
> Best RegardsDaniele
> -- 
> Dr. Daniele Riggi, PhD student
> University of Milano-Bicocca
> Department of Statistics
> Building U7, Via Bicocca degli Arcimboldi, 8
> 20126 Milano, Italy
> cell. +39 328 3380690
> mailto: daniele.riggi at gmail.com
> 	[[alternative HTML version deleted]]
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

*** --- ***
Christian Hennig
University College London, Department of Statistical Science
Gower St., London WC1E 6BT, phone +44 207 679 1698
chrish at stats.ucl.ac.uk, www.homepages.ucl.ac.uk/~ucakche

More information about the R-help mailing list