R help - Feb 2014 - update.formula() to add interaction terms

Hi,
I have a list of formulae that I need to modify. For each formula, I need
to add an interaction term between the first independent variable and the
last one. I want to write a function to achieve that because the list of
formulae to modify is long.

For example, if the formula is y ~ x1 + x2 + x3 + x4, then I need to turn
it either into

y ~ x1*x4 + x2 + x3

or

y ~ x1 + x2 + x3 + x4 + x4:x1

(I know they are the same, but one may be easier to work with than the
other).


Suppose I have the formula a which is defined as a <- formula(y ~
x1+x2+x3+x4)

I know how to access to the first and last independent variables of my
formula:

firstvar <- all.vars(a[[3]])[1]
lastvar <- all.vars(a[[3]])[length( all.vars(a[[3]]))]


What I can't figure out is how to use update.formula() to include my
interaction term in order to get y ~ x1+x2+x3+x4 + x1:x4. Specifically,

update.formula(a, ~ . + paste(all.vars(a[[3]])[1],firstvar, lastvar, sep
':'))

is not producing y ~ x1+x2+x3+x4 + x1:x4.

Any idea?

Thanks in advance,
MP

	[[alternative HTML version deleted]]

Is this the sort of thing you are looking for?
   > fm <- y ~ x1 + x2 + x3 + log(x4)
   > # Use terms() instead of just all.vars() to keep log(x4) as log(x4)
   > xVars <- with(attributes(terms(fm)),
as.list(variables)[-c(1,response+1)])
   > str(xVars)
   List of 4
    $ : symbol x1
    $ : symbol x2
    $ : symbol x3
    $ : language log(x4)
   > # use bquote to make the addition to the formula
   > update(fm, bquote( ~ . + .(xVars[[1]]) * .(xVars[[length(xVars)]])))
   y ~ x1 + x2 + x3 + log(x4) + x1:log(x4)

As a function it would be
   addInteraction <- function(formula){
      xVars <- with(attributes(terms(formula)),
as.list(variables)[-c(1,response+1)])
     update(formula, bquote( ~ . + .(xVars[[1]]) * .(xVars[[length(xVars)]])))
   }
used as
   > addInteraction(y~x1+x2+sqrt(x3))
   y ~ x1 + x2 + sqrt(x3) + x1:sqrt(x3)

If the last 'term' in the formula is a compound like x4:x5 and you
want x1:x4:x5 added you will need to do more work (look at the
'factors' attribute of terms()'s output) - currently it adds x1:x5.


Bill Dunlap
TIBCO Software
wdunlap tibco.com

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org] On Behalf
> Of Marie-Pierre Sylvestre
> Sent: Monday, February 03, 2014 12:54 PM
> To: r-help at r-project.org
> Subject: [R] update.formula() to add interaction terms
> 
> Hi,
> I have a list of formulae that I need to modify. For each formula, I need
> to add an interaction term between the first independent variable and the
> last one. I want to write a function to achieve that because the list of
> formulae to modify is long.
> 
> For example, if the formula is y ~ x1 + x2 + x3 + x4, then I need to turn
> it either into
> 
> y ~ x1*x4 + x2 + x3
> 
> or
> 
> y ~ x1 + x2 + x3 + x4 + x4:x1
> 
> (I know they are the same, but one may be easier to work with than the
> other).
> 
> 
> Suppose I have the formula a which is defined as a <- formula(y ~
> x1+x2+x3+x4)
> 
> I know how to access to the first and last independent variables of my
> formula:
> 
> firstvar <- all.vars(a[[3]])[1]
> lastvar <- all.vars(a[[3]])[length( all.vars(a[[3]]))]
> 
> 
> What I can't figure out is how to use update.formula() to include my
> interaction term in order to get y ~ x1+x2+x3+x4 + x1:x4. Specifically,
> 
> update.formula(a, ~ . + paste(all.vars(a[[3]])[1],firstvar, lastvar, sep
> ':'))
> 
> is not producing y ~ x1+x2+x3+x4 + x1:x4.
> 
> Any idea?
> 
> Thanks in advance,
> MP
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> 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.