[R] normal mixture model

[Christian Hennig]

Hi Cindy,

you need the summary function
> mclustsummary <- summary(mclustBICoutputobject,data)
to get all the information. Some (like best model) is given if you just 
print out the summary object. Some other information (like 
estimated parameter values) are accessible as components of the summary 
object, like
mclustsummary$parameters$...
Try
> str(mclustsummary)
to see what's there (unfortunately this is not fully documented).

For more detail see the help pages.

Hope this helps,
Christian

On Sun, 26 Jul 2009, cindy Guo wrote:

> Hi, Christian,
>
> Thank you for the reply. I just tried. Does the function mclustBIC only give
> the best model, or does it also do EM to get the cluster means and variances
> according to the best model it picks? I didn't find it.  Is there a way to
> automatically select the best number of components and do EM? Because I need
> to do the normal mixture model in a loop (one EM at an iteration), so I want
> it to do everything automatically.
> Thanks,
>
> Cindy
>
> On Sun, Jul 26, 2009 at 3:46 PM, Christian Hennig <chrish at stats.ucl.ac.uk>wrote:
>
>> You can use mclustBIC in package mclust (uses the BIC for deciding about
>> the number of components and hierarchical clustering for initialisation).
>>
>> Christian
>>
>>
>> On Sun, 26 Jul 2009, cindy Guo wrote:
>>
>>   Hi, All,
>>>
>>> I want to fit a normal mixture model. Which package in R is best for this?
>>> I
>>> was using the package 'mixdist', but I need to group the data into groups
>>> before fitting model, and different groupings seem to lead to different
>>> results. What other package can I use which is stable? And are there
>>> packages that can automatically determine the number of components?
>>>
>>> Thank you,
>>>
>>> Cindy
>>>
>>>        [[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<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
>>
>

*** --- ***
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