What error distribution should I use?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

What error distribution should I use?

Lívia Dorneles Audino
I'm trying to make a glmm to identify the relationship between insect
species richness with fragment size, isolation and time (different years).
I already tried to analyse it using poisson distribution error, but I
always face with the following warning:
*glm.fit: fitted probabilities numerically 0 or 1 occurred *

This is probably hapenning because my dataset has a lot of zeros. So, what
error distribution should I use?

--
*Lívia *

        [[alternative HTML version deleted]]


______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

chuck.01
Could you please post a small example of your data and code which gives you this error.  Your assumed error distribution sounds reasonable.  I am interested as to why you have zeros... you have sites with species richness ==0 ??    


Lívia Dorneles Audino wrote
I'm trying to make a glmm to identify the relationship between insect
species richness with fragment size, isolation and time (different years).
I already tried to analyse it using poisson distribution error, but I
always face with the following warning:
*glm.fit: fitted probabilities numerically 0 or 1 occurred *

This is probably hapenning because my dataset has a lot of zeros. So, what
error distribution should I use?

--
*Lívia *

        [[alternative HTML version deleted]]


______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

bbolker
In reply to this post by Lívia Dorneles Audino
Lívia Dorneles Audino <livia.audino <at> gmail.com> writes:

>
> I'm trying to make a glmm to identify the relationship between insect
> species richness with fragment size, isolation and time (different years).
> I already tried to analyse it using poisson distribution error, but I
> always face with the following warning:
> *glm.fit: fitted probabilities numerically 0 or 1 occurred *
>
> This is probably hapenning because my dataset has a lot of zeros. So, what
> error distribution should I use?
>

  I know you haven't gotten a lot of help on r-sig-mixed-models (sorry),
but it would probably be better to post this question there.  The answer
is that this is a warning, not an error, so it indicates a need for
caution but not necessarily that anything is wrong.  In this case,
an internal call to glm.fit() has difficulty when it tries to fit
a subset of that data that are all-zero or all-one.  It's quite possibly
OK, provided that you've looked at your results, plotted predicted values,
etc., and everything seems to make sense.

______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

Peter Ehlers
On 2012-03-27 15:11, Ben Bolker wrote:

> Lívia Dorneles Audino<livia.audino<at>  gmail.com>  writes:
>
>>
>> I'm trying to make a glmm to identify the relationship between insect
>> species richness with fragment size, isolation and time (different years).
>> I already tried to analyse it using poisson distribution error, but I
>> always face with the following warning:
>> *glm.fit: fitted probabilities numerically 0 or 1 occurred *
>>
>> This is probably hapenning because my dataset has a lot of zeros. So, what
>> error distribution should I use?
>>
>
>    I know you haven't gotten a lot of help on r-sig-mixed-models (sorry),
> but it would probably be better to post this question there.  The answer
> is that this is a warning, not an error, so it indicates a need for
> caution but not necessarily that anything is wrong.  In this case,
> an internal call to glm.fit() has difficulty when it tries to fit
> a subset of that data that are all-zero or all-one.  It's quite possibly
> OK, provided that you've looked at your results, plotted predicted values,
> etc., and everything seems to make sense.
>

Livia:
  You might also find this quite extensive recent post from Ted Harding
informative:

   https://stat.ethz.ch/pipermail/r-help/2012-March/307352.html

Peter Ehlers

______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

Lívia Dorneles Audino
In reply to this post by bbolker
Thank's a lot for your help Ben.
I already posted my question on r-sig-mixed-models! However you already
answered what I needed to know.

2012/3/27 Ben Bolker <[hidden email]>

> Lívia Dorneles Audino <livia.audino <at> gmail.com> writes:
>
> >
> > I'm trying to make a glmm to identify the relationship between insect
> > species richness with fragment size, isolation and time (different
> years).
> > I already tried to analyse it using poisson distribution error, but I
> > always face with the following warning:
> > *glm.fit: fitted probabilities numerically 0 or 1 occurred *
> >
> > This is probably hapenning because my dataset has a lot of zeros. So,
> what
> > error distribution should I use?
> >
>
>   I know you haven't gotten a lot of help on r-sig-mixed-models (sorry),
> but it would probably be better to post this question there.  The answer
> is that this is a warning, not an error, so it indicates a need for
> caution but not necessarily that anything is wrong.  In this case,
> an internal call to glm.fit() has difficulty when it tries to fit
> a subset of that data that are all-zero or all-one.  It's quite possibly
> OK, provided that you've looked at your results, plotted predicted values,
> etc., and everything seems to make sense.
>
> ______________________________________________
> [hidden email] 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.
>


--
*Lívia Dorneles Audino*
Doutoranda em Entomologia
Laboratório de Ecologia e Conservação de Invertebrados
Universidade Federal de Lavras

        [[alternative HTML version deleted]]


______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

Lívia Dorneles Audino
In reply to this post by Peter Ehlers
Thank's for your help Peter!
This post is really very informative. It helped a lot!

2012/3/27 Peter Ehlers <[hidden email]>

> On 2012-03-27 15:11, Ben Bolker wrote:
>
>> Lívia Dorneles Audino<livia.audino<at>  gmail.com>  writes:
>>
>>
>>> I'm trying to make a glmm to identify the relationship between insect
>>> species richness with fragment size, isolation and time (different
>>> years).
>>> I already tried to analyse it using poisson distribution error, but I
>>> always face with the following warning:
>>> *glm.fit: fitted probabilities numerically 0 or 1 occurred *
>>>
>>> This is probably hapenning because my dataset has a lot of zeros. So,
>>> what
>>> error distribution should I use?
>>>
>>>
>>   I know you haven't gotten a lot of help on r-sig-mixed-models (sorry),
>> but it would probably be better to post this question there.  The answer
>> is that this is a warning, not an error, so it indicates a need for
>> caution but not necessarily that anything is wrong.  In this case,
>> an internal call to glm.fit() has difficulty when it tries to fit
>> a subset of that data that are all-zero or all-one.  It's quite possibly
>> OK, provided that you've looked at your results, plotted predicted values,
>> etc., and everything seems to make sense.
>>
>>
> Livia:
>  You might also find this quite extensive recent post from Ted Harding
> informative:
>
>  https://stat.ethz.ch/**pipermail/r-help/2012-March/**307352.html<https://stat.ethz.ch/pipermail/r-help/2012-March/307352.html>
>
> Peter Ehlers
>
>
> ______________________________**________________
> [hidden email] mailing list
> 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.
>


--
*Lívia Dorneles Audino*
Doutoranda em Entomologia
Laboratório de Ecologia e Conservação de Invertebrados
Universidade Federal de Lavras

        [[alternative HTML version deleted]]


______________________________________________
[hidden email] 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.
Reply | Threaded
Open this post in threaded view
|

Re: What error distribution should I use?

Lívia Dorneles Audino
In reply to this post by chuck.01
Hi Chuck!

Actually my data has a lot of zeros because I counted the number of species
that is exclusive to each studied fragment. So, my response variable is the
number of exclusive species. That's why it has a lot of zeros. I have
fragments that do not present exclusive species.

Thank's for your help!

2012/3/27 chuck.01 <[hidden email]>

> Could you please post a small example of your data and code which gives you
> this error.  Your assumed error distribution sounds reasonable.  I am
> interested as to why you have zeros... you have sites with species richness
> ==0 ??
>
>
>
> Lívia Dorneles Audino wrote
> >
> > I'm trying to make a glmm to identify the relationship between insect
> > species richness with fragment size, isolation and time (different
> years).
> > I already tried to analyse it using poisson distribution error, but I
> > always face with the following warning:
> > *glm.fit: fitted probabilities numerically 0 or 1 occurred *
> >
> > This is probably hapenning because my dataset has a lot of zeros. So,
> what
> > error distribution should I use?
> >
> > --
> > *Lívia *
> >
> >       [[alternative HTML version deleted]]
> >
> >
> > ______________________________________________
> > R-help@ 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.
> >
>
>
> --
> View this message in context:
> http://r.789695.n4.nabble.com/What-error-distribution-should-I-use-tp4509479p4510351.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> [hidden email] 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.
>


--
*Lívia Dorneles Audino*
Doutoranda em Entomologia
Laboratório de Ecologia e Conservação de Invertebrados
Universidade Federal de Lavras

        [[alternative HTML version deleted]]


______________________________________________
[hidden email] 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.