[R] how to define a geometric distribution for "glm"

[Amanda Li]

Hi Peter,

Thank you very much for your help! However, for my dataset, it may not
asymptotically work. May I ask whether you know how to define a new family?

Thank you very much again!

Best,
Amanda

2014-10-27 18:27 GMT-04:00 peter dalgaard <pdalgd at gmail.com>:

> The likelihood for the geometric distribution is the same as for the
> binomial distribution, except for the constant term, so estimates and LRT
> will be the same. The properties of the estimator will be different, e.g.
> the estimate of p is not unbiased, but asymptotically the likelihood
> procedures should work (asymptotic in this case means a reasonably large
> total number of both successes and failures, I suppose.)
>
> So, if your geometric variate is called y, with the R convention of
> counting the number of failures (not number of experiments), it should work
> with
>
> glm(cbind(1,y) ~ whatever, family="binomial")
>
> [The likelihood equivalence is fairly well-known in statistical theory as
> a counterargument to the strong likelihood principle that all inference
> should be based solely on the likelihood function.]
>
> - Peter D.
>
> > On 27 Oct 2014, at 22:29 , Amanda Li <amandali at uchicago.edu> wrote:
> >
> > Hello,
> >
> > I was trying to apply "glm" to a dataset that assumes geometric
> > distribution. I cannot use "glm.nb" in MASS package (negative.binomial
> (1))
> > because it tries to estimate this "1" while I am interested in "p", the
> > probability of success. Does anyone know how I can define a geometric
> > distribution within "family" so that I can use glm assuming geometric
> > distribution to estimate "p"?
> >
> > I am not sure how "quasi" within the family works in this case and I am
> not
> > sure whether it can be used to assume geometric distribution.
> >
> > Thanks in advance for your help! I really appreciate it!
> > Best regards,
> > Amanda
> >
> >       [[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.
>
> --
> Peter Dalgaard, Professor,
> Center for Statistics, Copenhagen Business School
> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
> Phone: (+45)38153501
> Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com
>
>
>
>
>
>
>
>
>

	[[alternative HTML version deleted]]