Displaying 3 results from an estimated 3 matches for "glim4".
Did you mean:
glim
2002 Jul 24
0
R code for Frequency and Count Data
...Count data. Oxford University Press (1995)
on my web page
www.luc.ac.be/~jlindsey
There are over 160 data sets, about one-half of which have the R
analysis output (the examples), the rest being for the exercises.
Doing this has reinforced my conclusion that R's glm simply does not
stack up to GLIM4 for generalised linear models. I guess that
generalist software just cannot compete with the specialists.
Jim
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "he...
2000 Dec 18
3
problems with glm (PR#771)
...st problem: with diagonal values weighted out in a transition
matrix (mover-stayer model), glm wrongly estimates the diagonals to be
the observed plus 0.1. This was correct in 90.1 a year ago but already
wrong in 1.0.1. I don't have any versions in between installed.
> # correct values from GLIM4
> # unit observed fitted residual
> # (1) 118 1.586 0.000
> # 2 14 11.836 0.629
> # 3 8 8.892 -0.299
> # 4 12 13.272 -0.349
> # 5 12 12.758 -0.212
> # (6) 212...
2000 Dec 18
3
problems with glm (PR#771)
...st problem: with diagonal values weighted out in a transition
matrix (mover-stayer model), glm wrongly estimates the diagonals to be
the observed plus 0.1. This was correct in 90.1 a year ago but already
wrong in 1.0.1. I don't have any versions in between installed.
> # correct values from GLIM4
> # unit observed fitted residual
> # (1) 118 1.586 0.000
> # 2 14 11.836 0.629
> # 3 8 8.892 -0.299
> # 4 12 13.272 -0.349
> # 5 12 12.758 -0.212
> # (6) 212...