Nick Cox wrote:
Historically the abbreviation -glim- (or more
precisely GLIM) was used for software implementing
generalized l.m.s (sensu Nelder and Wedderburn).
My guess is that -glm- was used in Stata as a
slightly different name precisely to avoid any
inference (implication, too) that -glm- matched
GLIM one-to-one, which it certainly did not.
But if Joe Hilbe is watching he can beat guessing
and tell us why he chose that name.
--------------------------------------------------------------------------------
I thought that users of other statistical software packages (SAS and SPSS
excepted) also use GLM to stand for generalized linear model. For example,
the quote below is excerpted from a post to the S-Plus list. I've seen the
initialism used in this manner on GenStat's list, as well, and I believe
even by Sir John Nelder posting there. At least one source
( www.ucalgary.ca/infoserve/Vol8.6/glim.html ) states GLIM to stand for
Generalised Linear Interactive Modeling.
And while we're at it, just what is a general linear model, anyway? I've
always understood it as SAS's commercial (marketing) term for "linear
model," that is, a distinction without a difference. Am I missing
something?
Joseph Coveney
"As you are not in the SAS world here, be aware that the G in glm and gam is
`generalized' not `general' (SAS's GLM merely highlights that for many years
it was incapable of fitting a linear model with factors and continuous
variables)."
(Prof. Brian D. Ripley, posted May 29, 1999, archived at
www.biostat.wustl.edu/archives/ html/s-news/1999-05/msg00321.html )
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
