package to fit mixtures of student-t distributions

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

package to fit mixtures of student-t distributions

R help mailing list-2
Hello!

I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).

Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.

I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).

The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.

This is in a nutshell what I want the software to accomplish:

Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.

I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.

Thanks!

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

Bert Gunter-2
Offlist, because this is (a) an opinion and (b) about statistics and
therefore offtopic.

I don't know whether any such package exists, but I would predict that
this is likely to be overdetermined (too many parameters) and
therefore unlikely to be a successful strategy. Fitting a mixture of
Gaussians is already difficult enough.

Feel free to ignore, of course, and no need to reply.

Cheers,
Bert


Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )


On Thu, Jun 29, 2017 at 5:41 AM, vare vare via R-help
<[hidden email]> wrote:

> Hello!
>
> I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).
>
> Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.
>
> I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).
>
> The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.
>
> This is in a nutshell what I want the software to accomplish:
>
> Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.
>
> I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.
>
> Thanks!
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

Ranjan Maitra-3
In reply to this post by R help mailing list-2
Would package "teigen" help?

Ranjan

On Thu, 29 Jun 2017 14:41:34 +0200 vare vare via R-help <[hidden email]> wrote:

> Hello!
>
> I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).
>
> Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.
>
> I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).
>
> The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.
>
> This is in a nutshell what I want the software to accomplish:
>
> Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.
>
> I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.
>
> Thanks!
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

--
Important Notice: This mailbox is ignored: e-mails are set to be deleted on receipt. Please respond to the mailing list if appropriate. For those needing to send personal or professional e-mail, please use appropriate addresses.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

R help mailing list-2
In reply to this post by Bert Gunter-2
I don’t see how neither a) or b) applies to this question nor the technical merit of the remark about  mixture models.

Do you have a suggestion for a more appropriate forum for this issue/question? (stackoverflow basically sent me here).

Kind regards

> On 29. Jun 2017, at 16:58, Bert Gunter <[hidden email]> wrote:
>
> Offlist, because this is (a) an opinion and (b) about statistics and
> therefore offtopic.
>
> I don't know whether any such package exists, but I would predict that
> this is likely to be overdetermined (too many parameters) and
> therefore unlikely to be a successful strategy. Fitting a mixture of
> Gaussians is already difficult enough.
>
> Feel free to ignore, of course, and no need to reply.
>
> Cheers,
> Bert
>
>
> Bert Gunter
>
> "The trouble with having an open mind is that people keep coming along
> and sticking things into it."
> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
>
>
> On Thu, Jun 29, 2017 at 5:41 AM, vare vare via R-help
> <[hidden email]> wrote:
>> Hello!
>>
>> I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).
>>
>> Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.
>>
>> I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).
>>
>> The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.
>>
>> This is in a nutshell what I want the software to accomplish:
>>
>> Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.
>>
>> I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.
>>
>> Thanks!
>>
>> ______________________________________________
>> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
>> 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.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

mark leeds
In reply to this post by Ranjan Maitra-3
Hi: The R package below may be of use to you.

https://journal.r-project.org/archive/2009-1/RJournal_2009-1_Ardia+et+al.pdf


On Thu, Jun 29, 2017 at 12:15 PM, Ranjan Maitra <[hidden email]> wrote:

> Would package "teigen" help?
>
> Ranjan
>
> On Thu, 29 Jun 2017 14:41:34 +0200 vare vare via R-help <
> [hidden email]> wrote:
>
> > Hello!
> >
> > I am new to R (before used python exclusively and would actually call
> the R solution for this issue inside a python notebook, hope that doesn’t
> disqualify me right of the batch).
> >
> > Right now I am  looking for a piece of software  to fit a 1D data sample
> to a mixture of t-distributions.
> >
> > I searched quite a while already and it seems to be that this is a
> somehwat obscure endeavor as most search results turn up for mixture of
> gaussians (what I am not interested here).
> >
> > The most promising candidates so far are the "AdMit" and "MitSEM" R
> packages. However I do not know R and find the description of these
> packages rather comlple and it seems their core objective is not the
> fitting of mixtures of t’s but instead use this as a step to accomplish
> something else.
> >
> > This is in a nutshell what I want the software to accomplish:
> >
> > Fitting a mixture of t-distributions to some data and estimate the
> "location" "scale" and "degrees of freedom" for each.
> >
> > I hope someone can point me to a simple package, I can’t believe that
> this is such an obscure use case.
> >
> > Thanks!
> >
> > ______________________________________________
> > [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> > 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.
>
> --
> Important Notice: This mailbox is ignored: e-mails are set to be deleted
> on receipt. Please respond to the mailing list if appropriate. For those
> needing to send personal or professional e-mail, please use appropriate
> addresses.
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

Ingmar Visser
In reply to this post by R help mailing list-2
gamlss.mx can fit these I believe (although no experience with these myself)
flexmix may be (relatively easily) adaptable to accomplish this as well
hth, Ingmar

Ingmar Visser
Universitair Hoofddocent ontwikkelingspsychologie | Directeur College
Psychologie
Afdeling Psychologie | Faculteit Maatschappij- en Gedragswetenschappen |
Universiteit van Amsterdam
Bezoek | Nieuwe Achtergracht 129B | Kamer G 1.18
Post | Postbus 15933 | 1001 NK Amsterdam
Pakketpost | Valckenierstraat 59 | 1018 XE Amsterdam
T: +31205256723 | M: +31647260824 | e: [hidden email]

On Thu, Jun 29, 2017 at 2:41 PM, vare vare via R-help <[hidden email]>
wrote:

> Hello!
>
> I am new to R (before used python exclusively and would actually call the
> R solution for this issue inside a python notebook, hope that doesn’t
> disqualify me right of the batch).
>
> Right now I am  looking for a piece of software  to fit a 1D data sample
> to a mixture of t-distributions.
>
> I searched quite a while already and it seems to be that this is a
> somehwat obscure endeavor as most search results turn up for mixture of
> gaussians (what I am not interested here).
>
> The most promising candidates so far are the "AdMit" and "MitSEM" R
> packages. However I do not know R and find the description of these
> packages rather comlple and it seems their core objective is not the
> fitting of mixtures of t’s but instead use this as a step to accomplish
> something else.
>
> This is in a nutshell what I want the software to accomplish:
>
> Fitting a mixture of t-distributions to some data and estimate the
> "location" "scale" and "degrees of freedom" for each.
>
> I hope someone can point me to a simple package, I can’t believe that this
> is such an obscure use case.
>
> Thanks!
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

David Winsemius
In reply to this post by R help mailing list-2

> On Jun 29, 2017, at 10:02 AM, vare vare via R-help <[hidden email]> wrote:
>
> I don’t see how neither a) or b) applies to this question nor the technical merit of the remark about  mixture models.
>
> Do you have a suggestion for a more appropriate forum for this issue/question?

CrossValidated.com, although if you frame it in terms of "I need a language specific response", they often will refer questioners back to StackOverflow. There seems to be a widespread aversion to doing people's searching for them. You should read the Posting Guide and the help pages for any online forum you consider posting to. SO has a filter that prevents LMGTFY responses, but Rhelp has no such prohibition.

http://lmgtfy.com/?q=package+to+fit+mixtures+of+student-t+distributions

> (stackoverflow basically sent me here).

I question whether "stackoverflow basically sent me here" in any meaningful sense. There is a response yesterday on StackOverflow from Ben Bolker in which he constructed a working example (which should have been provided by you) and he then illustrated the use of a function from the `teigen` package that seemed to work fairly well. There were not any comments that suggested you should use the rhelp mailing list. You have not commented on, upvoted, or accepted that answer using the mechanisms given to you by the SO people to recognize the answerer's effort or guide the "helping" process.  So one suggestion would be to:

-- show what searching you have done,  (Maybe you really do need methods advice?)

-- and why the suggestions so far are not satisfactory.

--
David.


>
> Kind regards
>
>> On 29. Jun 2017, at 16:58, Bert Gunter <[hidden email]> wrote:
>>
>> Offlist, because this is (a) an opinion and (b) about statistics and
>> therefore offtopic.
>>
>> I don't know whether any such package exists, but I would predict that
>> this is likely to be overdetermined (too many parameters) and
>> therefore unlikely to be a successful strategy. Fitting a mixture of
>> Gaussians is already difficult enough.
>>
>> Feel free to ignore, of course, and no need to reply.
>>
>> Cheers,
>> Bert
>>
>>
>> Bert Gunter
>>
>> "The trouble with having an open mind is that people keep coming along
>> and sticking things into it."
>> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
>>
>>
>> On Thu, Jun 29, 2017 at 5:41 AM, vare vare via R-help
>> <[hidden email]> wrote:
>>> Hello!
>>>
>>> I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).
>>>
>>> Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.
>>>
>>> I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).
>>>
>>> The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.
>>>
>>> This is in a nutshell what I want the software to accomplish:
>>>
>>> Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.
>>>
>>> I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.
>>>
>>> Thanks!
>>>
>>> ______________________________________________
>>> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
>>> 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.
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> 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.

David Winsemius
Alameda, CA, USA

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.
Reply | Threaded
Open this post in threaded view
|

Re: package to fit mixtures of student-t distributions

R help mailing list-2
In reply to this post by Ingmar Visser
Thanks Ranjan,

that provides exactly what I need.
I also got a more elaborate answer on stackoverflow  (after the same question got rejected form cross validated…) with a running example for this package:

https://stackoverflow.com/questions/44825529/package-to-fit-mixtures-of-student-t-distributions/44827220#44827220 <https://stackoverflow.com/questions/44825529/package-to-fit-mixtures-of-student-t-distributions/44827220#44827220>

Kind regards
Vare

> Would package "teigen" help?
>
> Ranjan


>
>
> On Thu, Jun 29, 2017 at 2:41 PM, vare vare via R-help <[hidden email] <mailto:[hidden email]>> wrote:
> Hello!
>
> I am new to R (before used python exclusively and would actually call the R solution for this issue inside a python notebook, hope that doesn’t disqualify me right of the batch).
>
> Right now I am  looking for a piece of software  to fit a 1D data sample to a mixture of t-distributions.
>
> I searched quite a while already and it seems to be that this is a somehwat obscure endeavor as most search results turn up for mixture of gaussians (what I am not interested here).
>
> The most promising candidates so far are the "AdMit" and "MitSEM" R packages. However I do not know R and find the description of these packages rather comlple and it seems their core objective is not the fitting of mixtures of t’s but instead use this as a step to accomplish something else.
>
> This is in a nutshell what I want the software to accomplish:
>
> Fitting a mixture of t-distributions to some data and estimate the "location" "scale" and "degrees of freedom" for each.
>
> I hope someone can point me to a simple package, I can’t believe that this is such an obscure use case.
>
> Thanks!
>
> ______________________________________________
> [hidden email] <mailto:[hidden email]> mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help <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.
>


        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
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.