R help - Nov 2004 - A GLM for a parameter of the error distribution.

Dear Sir/Madam,

 

I am an experienced programmer that has had little exposure to R so
please forgive me if I am inaccurate in places. I have an objective with
which I can not find a solution.  I would like to fit a GLM to binomial
data that includes r number of success from n trials.  The GLM that I am
proposing describes P, the binomial probability of success. The link
function proposed is the logit function thus:

 

log(Py,f,g,eu /(1- Py,f,g,eu)) = B0 + By + Bf + Bg + Beu

 

ry,f,g,eu ~ bin(ny,f,g,eu, Py,f,g,eu)

 

How would I go about fitting a GLM that describes a parameter of the
error distribution?

 

I would be most grateful of any assistance that you can offer me in this
matter.

Sorry to be a pain.  

 

Best regards,

 

Tom

 

 

 

Thomas Robert Carruthers

 

Renewable Resources Assessment Group

Royal School of Mines 

Department of Environmental Science and Technology 

South Kensington Campus

Imperial College

London

 

SW7 2AZ

UK

 

Tel: +44 (0)20 7589 5111 ext 59276

Mob: +44 07876 891589

Email: thomas.carruthers@imperial.ac.uk

 

 

	[[alternative HTML version deleted]]