[R] glm - poisson

ronggui 0034058 at fudan.edu.cn
Fri Mar 18 17:23:57 CET 2005


i think that is ok.
"when a response count n(i) has index equal to t(i),the sample rate is n(i)/t(i),its expected value is u(i)/t(i)......a loglinear model for theexpeted rate has form log(u(i)/t(i))=a+bx"(agresti,2002)
this model can use glm-poisson with a offset term to estimate

On Thu, 17 Mar 2005 15:22:02 -0800 (PST)
Oli <olilili at yahoo.com> wrote:

> Hello,
> 
> I have a question on glm - poisson. I would like to
> fit a model on proportion of counts with some factors,
> and I decided to use glm - poisson because count data
> is essentially related to poisson. However, since I
> need to deal with proportion, and glm - poisson does
> not allow non-integer responses, I am thinking of
> adding an offset term = log(total). Is it the right
> approach to it? or should I use quasi-poisson family,
> or should I even use weights?
> 
> Thanks for your help!
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html




More information about the R-help mailing list