R help - Jan 2005 - glm fit with no intercept

Dear R-help list members,

I am currently trying to fit a generalized linear model using a binomial
with the canonical link. The usual solution is to use the R function glm()
in the package "stats". However, I run into problem when I want to fit
a
glm without an intercept. It is indicated that the solution is in changing
the function glm.fit (also in "stats"), by specifying intercept=FALSE.
I
have not been successful in getting any output though.

Any suggestion on how to fit a glm with no intercept?

Regards,
Gudrun

-- 
Gudrun Jonasdottir, M.Sc.
Matematiska institutionen
Stockholms Universitet
SE- 106 91 Stockholm

Work: +46 (0)8 16 45 56
Mobile: +46 (0)709 779 800
--------------------------------------------
"N?r inget annat hj?lper, l?s instruktionsboken."
- Canns axiom

Hi Gudrun,

try the following:

y <- rbinom(100, 1, 0.5)
x <- rnorm(100)
##############
glm(y~x, family=binomial)
glm(y~x-1, family=binomial)

I hope it helps.

Best,
Dimitris

----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven

Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/16/336899
Fax: +32/16/337015
Web: http://www.med.kuleuven.ac.be/biostat
     http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm


----- Original Message ----- 
From: "Gudrun Jonasdottir" <gudrunj at math.su.se>
To: <r-help at stat.math.ethz.ch>
Sent: Friday, January 07, 2005 3:40 PM
Subject: [R] glm fit with no intercept

> Dear R-help list members,
>
> I am currently trying to fit a generalized linear model using a 
> binomial
> with the canonical link. The usual solution is to use the R function 
> glm()
> in the package "stats". However, I run into problem when I want
to
> fit a
> glm without an intercept. It is indicated that the solution is in 
> changing
> the function glm.fit (also in "stats"), by specifying 
> intercept=FALSE. I
> have not been successful in getting any output though.
>
> Any suggestion on how to fit a glm with no intercept?
>
> Regards,
> Gudrun
>
> -- 
> Gudrun Jonasdottir, M.Sc.
> Matematiska institutionen
> Stockholms Universitet
> SE- 106 91 Stockholm
>
> Work: +46 (0)8 16 45 56
> Mobile: +46 (0)709 779 800
> --------------------------------------------
> "N?r inget annat hj?lper, l?s instruktionsboken."
> - Canns axiom
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
>

"Gudrun Jonasdottir" <gudrunj at math.su.se> writes:
> Dear R-help list members,
> 
> I am currently trying to fit a generalized linear model using a binomial
> with the canonical link. The usual solution is to use the R function glm()
> in the package "stats". However, I run into problem when I want
to fit a
> glm without an intercept. It is indicated that the solution is in changing
> the function glm.fit (also in "stats"), by specifying
intercept=FALSE. I
> have not been successful in getting any output though.
> 
> Any suggestion on how to fit a glm with no intercept?
Use -1 in the model formula.

  x <- sample(0:1,10,rep=T)
  d <- runif(10)
  summary(glm(x~d-1,family=binomial))
> "N?r inget annat hj?lper, l?s instruktionsboken."
> - Canns axiom
.. i n?rvarande fall: "An Introduction to R", Ch.11.


-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907

As in any modelling situation in R or S, to suppress the
intercept you simply put a ``-1'' in the formula.  E.g.:

	fit <- glm(y~x-1,family=binomial)

			cheers,

				Rolf Turner
				rolf at math.unb.ca

On Fri, 7 Jan 2005, Gudrun Jonasdottir wrote:
> I am currently trying to fit a generalized linear model using a binomial
> with the canonical link. The usual solution is to use the R function glm()
> in the package "stats". However, I run into problem when I want
to fit a
> glm without an intercept. It is indicated that the solution is in changing
Where is it so indicated?
> the function glm.fit (also in "stats"), by specifying
intercept=FALSE. I
> have not been successful in getting any output though.
>
> Any suggestion on how to fit a glm with no intercept?
>From MASS4:
ldose <- rep(0:5, 2)
numdead <- c(1, 4, 9, 13, 18, 20, 0, 2, 6, 10, 12, 16)
sex <- factor(rep(c("M", "F"), c(6, 6)))
SF <- cbind(numdead, numalive = 20 - numdead)
glm(SF ~ sex+ ldose, family = binomial)
glm(SF ~ sex + ldose - 1, family = binomial) # no intercept
glm(SF ~ ldose - 1, family = binomial) # and not even a factor

Please tell us more of what you have tried.

-- 
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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595