R help - Jul 2010 - nls + quasi-poisson distribution

Hello R-helpers,

I would like to fit a non-linear function to data (Discrete X axis,  
over-dispersed Poisson values on the Y axis).

I found the functions gnlr in the gnlm package from Jim Lindsey: this can  
handle nonlinear regression equations for the parameters of Poisson and  
negative binomial distributions, among others. I also found the function  
nls2 in the software package accompanying the book "Statistical tools for  
nonlinear regression" by Huet et al: this can handle nonlinear regression  
with Poisson distributed Y-axis values.

I was wondering if there was any other option: specifically, any option  
that handled nonlinear fitting with quasi-Poisson distributions (to handle  
the overdispersion).

This is a very new area for me, and I am still trying to figure out the  
best way to do this, so I would appreciate any and all pointers.

Thanks much, Suresh

Dear Suresh,

The gnm package for generalized nonlinear models might be what you want
here. This allows you to specify nonlinear models with family=quasipoisson.

For an introduction to the package see the article in R News:
http://cran.r-project.org/doc/Rnews/Rnews_2007-2.pdf

If your model requires a custom "nonlin" function and you get stuck on
this, or you have any other queries about the package, I'm happy to be
contacted directly.

Best regards,

Heather


On 06/07/10 07:04, Suresh Krishna wrote:
> 
> Hello R-helpers,
> 
> I would like to fit a non-linear function to data (Discrete X axis,
> over-dispersed Poisson values on the Y axis).
> 
> I found the functions gnlr in the gnlm package from Jim Lindsey: this
> can handle nonlinear regression equations for the parameters of Poisson
> and negative binomial distributions, among others. I also found the
> function nls2 in the software package accompanying the book
"Statistical
> tools for nonlinear regression" by Huet et al: this can handle
nonlinear
> regression with Poisson distributed Y-axis values.
> 
> I was wondering if there was any other option: specifically, any option
> that handled nonlinear fitting with quasi-Poisson distributions (to
> handle the overdispersion).
> 
> This is a very new area for me, and I am still trying to figure out the
> best way to do this, so I would appreciate any and all pointers.
> 
> Thanks much, Suresh
> 
> ______________________________________________
> 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.