Dear all,
I am implementing a redundant variables F-test. For that I need to
compute 2 models, restricted and unrestricted, then extracting the
residuals to calculate the test statistic.
I borrowed this elegant solution from the LMtest package to rebuild the
first of the matrices involved (the unrestricted one) on the basis of
model spec. and data
>red.var.test<-function(formula, red.vars, data = list()) {
> mf <- model.frame(formula, data = data)
> y <- model.response(mf)
> modelterms <- terms(formula, data = data)
> X <- model.matrix(modelterms, data = data) # unrestricted model
matrix
>
> n <- nrow(X)
> k <- ncol(X)
>
...but then I had to resort to this solution of mine to select the
restricted m.m. according to the possibly redundant regressors specified
in the character vector red.vars
> Z<-X[,!(dimnames(X)[[2]]%in%red.vars)] # restricted model matrix
>
and then the rest...
> q<-dim(X)[[2]]-dim(Z)[[2]]
>
> umod<-lm(y~X)
> rmod<-lm(y~Z)
>
> ures<-umod$resid
> rres<-rmod$resid
>
> URSS<-sum(ures^2)
> RRSS<-sum(rres^2)
>
> test <- ((RRSS - URSS)/q) / (URSS / (n-k))
> (...) }
Now to my question: the above works just fine when the names in red.vars
are *exactly* those of regressors in the model frame, e.g. if I am
taking logs of one variable, say, "a", in model mod<-lm(d ~ log(a) + b +
c), testing joint redundancy of a and b I have to write
red.var.test(mod, c("log(a)","b"))
Is there a generic way to restrict the model matrix, so as to write only
the names of regressors involved without having to bother about the
transformations, such as
red.var.test(mod, c("a","b"))
maybe in the style of the first rows?
Thank you in advance for your insights
Giovanni
Giovanni Millo
R&D Dept.
Assicurazioni Generali SpA
Trieste, Italy
Ai sensi del D.Lgs. 196/2003 si precisa che le informazioni contenute in questo
messaggio sono riservate ed a uso esclusivo del destinatario. Qualora il messaggio in
parola Le fosse pervenuto per errore, La invitiamo ad eliminarlo senza copiarlo e a
non inoltrarlo a terzi, dandocene gentilmente comunicazione. Grazie.<BR><BR>Pursuant
to Legislative Decree No. 196/2003, you are hereby informed that this message contains
confidential information intended only for the use of the addressee. If you are not
the addressee, and have received this message by mistake, please delete it and
immediately notify us. You may not copy or disseminate this message to anyone. Thank
you.
[[alternative HTML version deleted]]
______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
