Negative binomial parameterisation in mgcv
On 13/11/2013 09:38, Mark Payne wrote:
Dear R-help, The negative binomial distribution has several different parameterisations, but I can't seem to figure out what the exact one used in mgcv's negbin family is? negbin() requires a theta argument, but its not clear anywhere in the documentation (that I can find), how this parameter should be interpreted - for example, can it be less than 1?
I believe it is the same as MASS (the package) explained in MASS (the book). As its help says ....
Best wishes, MArk
Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595