Skip to content

starting with mixture distribution

2 messages · Alaios, Achim Zeileis

#
Hi.I was using the last night the fitdistr package to start with some fitting. I have some data sets that even though have a very gaussian distribution it looks like that also it has some very heavy tails, that can not be accurately be modelled by a gaussian distribution.Where should I give a try with mixture of distribution?
RegardsAlex
#
On Sun, 30 Nov 2014, Alaios via R-help wrote:

            
An overview of available distributions and corresponding fitting functions 
is in the "Distributions" task view. Finite mixture models and model-based 
clustering is discussed in the "Cluster" task view. See:
http://CRAN.R-project.org/view=Distributions
http://CRAN.R-project.org/view=Cluster