Dieter Menne
2009-Apr-01 12:10 UTC
[R] How to prevent inclusion of intercept in lme with interaction
Dear friends of lme,
After so many year with lme, I feel ashamed that I cannot get this to work.
Maybe it's a syntax problem, but possibly a lack of understanding.
We have growth curves of new dental bone that can well be modeled by a
linear growth curve, for two different treatments and several subjects as
random parameter. By definition, newbone is zero at t=0, so I tried to force
the curve through 0. In Pinheiro/Bates, this is done by including the -1
term, and it works well when treatment is not included (newbone~t-1), but
seems to have no effect in (newone ~ t*treat-1).
What's wrong? Some bracket missing? I tried a few variants.
Dieter Menne
#--------------------------------
library(nlme)
library(lattice)
# Generated data
set.seed(4711)
subject = as.factor(letters[1:5])
varslope = rnorm(length(subject),0,0.02)
cslope = c (0.1,0.15)
grd = expand.grid(t=seq(5,15,by=5),
subject=subject,treat=c("contr","test"))
grd$slope = varslope[grd$subject] + cslope[grd$treat]
grd$newbone = grd$slope*grd$t+rnorm(nrow(grd),0,0.2)
xyplot(newbone~t|treat,groups=subject,data=grd,
type="l",xlim=c(0,20),ylim=c(0,3))
# With intercept
grd.lme1 = lme(newbone~t*treat,data=grd,random=~1|subject)
grd$pred1 = predict(grd.lme1,level=0)
summary(grd.lme1)
# How go force intercept = 0 ???
grd.lme0 = lme(newbone~t*treat-1,data=grd,random=~1|subject)
grd$pred0 = predict(grd.lme0,level=0)
summary(grd.lme0)
# Gives true,
all.equal(grd$pred1,grd$pred0)
# Everything as expected without treat
grd.lme2 = lme(newbone~t,data=grd,random=~1|subject)
grd$pred2 = predict(grd.lme2,level=0)
summary(grd.lme2)
# Forced intercept = 0
grd.lme3 = lme(newbone~t-1,data=grd,random=~1|subject)
grd$pred3 = predict(grd.lme3,level=0)
summary(grd.lme3)
# As expected: not equal
all.equal(grd$pred2,grd$pred3)
#-------------------------------------------------------------------
R version 2.9.0 Under development (unstable) (2009-03-13 r48127)
i386-pc-mingw32
locale:
LC_COLLATE=German_Germany.1252;LC_CTYPE=German_Germany.1252;
LC_MONETARY=German_Germany.1252;LC_NUMERIC=C;LC_TIME=German_Germany.1252
attached base packages:
[1] stats graphics grDevices datasets utils methods base
other attached packages:
[1] lattice_0.17-20 nlme_3.1-90
loaded via a namespace (and not attached):
[1] grid_2.9.0 tools_2.9.0>
ONKELINX, Thierry
2009-Apr-01 12:35 UTC
[R] How to prevent inclusion of intercept in lme with interaction
Dear Dieter,
With t*treat the model allows for a different slope AND a different
intercept for each treatment. If you only want different slopes and all
intercepts equal to 0, then t:treat - 1 or t + t:treat - 1 is the model
you are looking for.
HTH,
Thierry
------------------------------------------------------------------------
----
ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature
and Forest
Cel biometrie, methodologie en kwaliteitszorg / Section biometrics,
methodology and quality assurance
Gaverstraat 4
9500 Geraardsbergen
Belgium
tel. + 32 54/436 185
Thierry.Onkelinx at inbo.be
www.inbo.be
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to
say what the experiment died of.
~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data.
~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of
data.
~ John Tukey
-----Oorspronkelijk bericht-----
Van: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
Namens Dieter Menne
Verzonden: woensdag 1 april 2009 14:11
Aan: r-help at stat.math.ethz.ch
Onderwerp: [R] How to prevent inclusion of intercept in lme with
interaction
Dear friends of lme,
After so many year with lme, I feel ashamed that I cannot get this to
work.
Maybe it's a syntax problem, but possibly a lack of understanding.
We have growth curves of new dental bone that can well be modeled by a
linear growth curve, for two different treatments and several subjects
as
random parameter. By definition, newbone is zero at t=0, so I tried to
force
the curve through 0. In Pinheiro/Bates, this is done by including the -1
term, and it works well when treatment is not included (newbone~t-1),
but
seems to have no effect in (newone ~ t*treat-1).
What's wrong? Some bracket missing? I tried a few variants.
Dieter Menne
#--------------------------------
library(nlme)
library(lattice)
# Generated data
set.seed(4711)
subject = as.factor(letters[1:5])
varslope = rnorm(length(subject),0,0.02)
cslope = c (0.1,0.15)
grd = expand.grid(t=seq(5,15,by=5),
subject=subject,treat=c("contr","test"))
grd$slope = varslope[grd$subject] + cslope[grd$treat]
grd$newbone = grd$slope*grd$t+rnorm(nrow(grd),0,0.2)
xyplot(newbone~t|treat,groups=subject,data=grd,
type="l",xlim=c(0,20),ylim=c(0,3))
# With intercept
grd.lme1 = lme(newbone~t*treat,data=grd,random=~1|subject)
grd$pred1 = predict(grd.lme1,level=0)
summary(grd.lme1)
# How go force intercept = 0 ???
grd.lme0 = lme(newbone~t*treat-1,data=grd,random=~1|subject)
grd$pred0 = predict(grd.lme0,level=0)
summary(grd.lme0)
# Gives true,
all.equal(grd$pred1,grd$pred0)
# Everything as expected without treat
grd.lme2 = lme(newbone~t,data=grd,random=~1|subject)
grd$pred2 = predict(grd.lme2,level=0)
summary(grd.lme2)
# Forced intercept = 0
grd.lme3 = lme(newbone~t-1,data=grd,random=~1|subject)
grd$pred3 = predict(grd.lme3,level=0)
summary(grd.lme3)
# As expected: not equal
all.equal(grd$pred2,grd$pred3)
#-------------------------------------------------------------------
R version 2.9.0 Under development (unstable) (2009-03-13 r48127)
i386-pc-mingw32
locale:
LC_COLLATE=German_Germany.1252;LC_CTYPE=German_Germany.1252;
LC_MONETARY=German_Germany.1252;LC_NUMERIC=C;LC_TIME=German_Germany.1252
attached base packages:
[1] stats graphics grDevices datasets utils methods base
other attached packages:
[1] lattice_0.17-20 nlme_3.1-90
loaded via a namespace (and not attached):
[1] grid_2.9.0 tools_2.9.0>
______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Dit bericht en eventuele bijlagen geven enkel de visie van de schrijver weer
en binden het INBO onder geen enkel beding, zolang dit bericht niet bevestigd is
door een geldig ondertekend document. The views expressed in this message
and any annex are purely those of the writer and may not be regarded as stating
an official position of INBO, as long as the message is not confirmed by a duly
signed document.
Christian Ritz
2009-Apr-01 12:37 UTC
[R] How to prevent inclusion of intercept in lme with interaction
Hi Dieter, the following model assumes a linear relationship between the response "newbone" and the independent variable "t" with a common intercept equal to 0 and treatment-dependent slopes: grd.lme0 <- lme(newbone~t:treat-1, data=grd, random=~1|subject) summary(grd.lme0) Christian
Dieter Menne
2009-Apr-01 12:40 UTC
[R] How to prevent inclusion of intercept in lme with interaction
ONKELINX, Thierry <Thierry.ONKELINX <at> inbo.be> writes:> With t*treat the model allows for a different slope AND a different > intercept for each treatment. If you only want different slopes and all > intercepts equal to 0, then t:treat - 1 or t + t:treat - 1 is the model > you are looking for.Thanks Thierry and Christian, I knew it was something stupid. Dieter