Dear list,
I am currently using the software gam.exact (with splus 2000),
written by Dominici et al, to run gam models to look at relationship
between air pollutants and mortality outcomes. The problem I have
is this:
Occasionally when I try to run the model with a certain number of
dependent variables, the model fails to run and comes up with the
following error message:
Errorin dim(data) <-dim:Length of data (99) doesn't match product
of dimensions (81).
There doesn't seem to be a particular pattern as to when the model
will or will not run. For example, the model may run with 10
variables whereas with 8 or 11 it will not. Similarly, it doesn't
appear dependent on the types of variables in the model.
Has anyone had any experience using gam.exact/can anyone
shed any light on the error message?
Any help greatly appreciated, thanks, Melanie
Dr Melanie Carder
Research statistician
Centre for Occupational and Environmental Health,
Faculty of Medicine, Dentistry, Nursing and Pharmacy,
The University of Manchester,
C Block, Level 4, Humanities Building,
Oxford Road,
Manchester M13 9PL
Tel:0161 275 5636
Fax: 0161 275 5595
|