[R] Generalized Estimating Functions
ripley@stats.ox.ac.uk
ripley at stats.ox.ac.uk
Mon May 13 19:09:46 CEST 2002
On Mon, 13 May 2002, Frederico Zanqueta Poleto wrote:
> Hi Thomas,
>
> Thank you for your help.
> You are right. I've tried the OME example and the working correlation was
> right.
> Maybe it is a problem with Poisson, what do you think?
No, that works too in other examples.
It's most likely a problem with your use of gee, as the working
correlation matrix shows, plus the fact that it only did one iteration.
Please check very carefully all your inputs.
> > desreg<-gee
> (desovas~ger+pop+fec+ger*pop+ger*fec+pop*fec,id=unidade,data=desovas.ovos.in
> viaveis,family=poisson,corstr="exchangeable")
> [1] "Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27"
> [1] "running glm to get initial regression estimate"
> [1] 0.93464328 0.02317831 0.29278443 0.21399704 0.11593571
> 0.05833780 -0.11513896
> > summary(desreg)
>
> GEE: GENERALIZED LINEAR MODELS FOR DEPENDENT DATA
> gee S-function, version 4.13 modified 98/01/27 (1998)
>
> Model:
> Link: Logarithm
> Variance to Mean Relation: Poisson
> Correlation Structure: Exchangeable
>
> Call:
> gee(formula = desovas ~ ger + pop + fec + ger * pop + ger * fec +
> pop * fec, id = unidade, data = desovas.ovos.inviaveis, family =
> poisson,
> corstr = "exchangeable")
>
> Summary of Residuals:
> Min 1Q Median 3Q Max
> -4.5893464 -1.5463047 -0.4124404 1.4106536 9.2329743
>
>
> Coefficients:
> Estimate Naive S.E. Naive z Robust S.E. Robust z
> (Intercept) 0.93464317 0.06012985 15.5437480 0.05954487 15.6964517
> ger 0.02317832 0.07001297 0.3310576 0.06915388 0.3351702
> pop 0.29278452 0.06565443 4.4594785 0.06505729 4.5004104
> fec 0.21399711 0.07252180 2.9507967 0.07150063 2.9929402
> ger:pop 0.11593572 0.07355902 1.5760911 0.07318636 1.5841166
> ger:fec 0.05833782 0.07000740 0.8333093 0.07054498 0.8269592
> pop:fec -0.11513904 0.07216442 -1.5955098 0.07149595 -1.6104274
>
> Estimated Scale Parameter: 1.366923
> Number of Iterations: 1
>
> Working Correlation
> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11] [,12]
> [1,] 1 0 0 0 0 0 0 0 0 0 0 0
> [2,] 0 0 0 0 0 0 0 0 0 0 0 0
> [3,] 0 0 0 0 0 0 0 0 0 0 0 0
> [4,] 0 0 0 0 0 0 0 0 0 0 0 0
> [5,] 0 0 0 0 0 0 0 0 0 0 0 0
> [6,] 0 0 0 0 0 0 0 0 0 0 0 0
> [7,] 0 0 0 0 0 0 0 0 0 0 0 0
> [8,] 0 0 0 0 0 0 0 0 0 0 0 0
> [9,] 0 0 0 0 0 0 0 0 0 0 0 0
> [10,] 0 0 0 0 0 0 0 0 0 0 0 0
> [11,] 0 0 0 0 0 0 0 0 0 0 0 0
> [12,] 0 0 0 0 0 0 0 0 0 0 0 0
>
>
> Best regards,
> --
> Frederico Zanqueta Poleto
> fred at poleto.com
> --
> "It would be possible to describe everything scientifically, but it would
> make no sense; it would be without meaning, as if you described a
> Beethoven symphony as a variation of wave pressure." Albert Einstein
>
>
> ------------- Original message follows -------------
>
> On Sun, 12 May 2002, Frederico Zanqueta Poleto wrote:
>
> > Hi,
> >
> > I'm trying to fit a marginal model via GEE but I'm getting strange
> > results and few problems.
> > If I set the working correlation as exchangeable I'm getting the same
> > fitting when I set as independent. Comparing to SAS results it shouldn't
> > happen.
> > If I try to use another working correlation (like AR-M or stat_M_dep), R
> > just exits without giving any error message.
>
> We probably need an example. The `OME' example in help(gee) gives
> different results using exchangeable and independence working
> correlations, so it isn't a universal problem.
>
> > Another doubt is how R estimates the scale parameter for Poisson? I get
> > 1.36 in R, 1.25 in SAS (estimation by the square root of DEVIANCE/DOF)
> > and 1.17 (estimation by the square root of Pearson's Chi-Square/DOF).
> > I'm using R 1.4.1 but all the problems apply for S-Plus 2000 too.
>
> This is because 1.36=1.17^2, I should think. The disperson parameter from
> gee() is a variance scale factor not a standard deviation scale factor.
>
> -thomas
>
> Thomas Lumley Asst. Professor, Biostatistics
> tlumley at u.washington.edu University of Washington, Seattle
>
>
>
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
>
--
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 272860 (secr)
Oxford OX1 3TG, UK Fax: +44 1865 272595
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
More information about the R-help
mailing list