R help - Sep 2011 - How to specify a variable name in the regression formula without hard coding it

I have a matrix called mat and y is the column number of my response and
x is a vector of the column numbers of my terms.  The variable name of y
can change, so I don't want to hardcode it.  I can find out the name as
follows:
>            names(mat)[y]
[1] "er12.l"

Then I can run the regression by hard coding the variable name as
follows:
>             mod <- try(rlm(er12.l ~ ., data=mat[zidx, c(y, x)]),
silent=TRUE)

But how would I do so without hard coding the name er12.l?

I set up a reproducible example.  In the following my regression formula
is aa ~ bb + cc, or more simply aa ~ .
How can I use the name of the first column without hard coding aa?

dat <- data.frame(aa=runif(50), bb=runif(50), cc=runif(50))
names(dat)[1]
lm(aa ~ ., data=dat[1:10,])

Thanks,

Roger
***************************************************************

This message is for the named person's use only. It may\...{{dropped:14}}

On Sep 8, 2011, at 9:03 AM, Bos, Roger wrote:
> I have a matrix called mat and y is the column number of my response  
> and
> x is a vector of the column numbers of my terms.  The variable name  
> of y
> can change, so I don't want to hardcode it.  I can find out the name  
> as
> follows:
>
>>           names(mat)[y]
> [1] "er12.l"
>
> Then I can run the regression by hard coding the variable name as
> follows:
>
>>            mod <- try(rlm(er12.l ~ ., data=mat[zidx, c(y, x)]),
> silent=TRUE)
>
> But how would I do so without hard coding the name er12.l?
>
> I set up a reproducible example.  In the following my regression  
> formula
> is aa ~ bb + cc, or more simply aa ~ .
> How can I use the name of the first column without hard coding aa?
>
> dat <- data.frame(aa=runif(50), bb=runif(50), cc=runif(50))
"dep <- names(dat)[1]; indep <- names(dat)[2]
> lm(aa ~ ., data=dat[1:10, c(dep,indep])
>

 > dep <- names(dat)[1]; indep <- names(dat)[2]
 > lm(aa ~ ., data=dat[1:10, c(dep,indep)])

Call:
lm(formula = aa ~ ., data = dat[1:10, c(dep, indep)])

Coefficients:
(Intercept)           bb
     0.58908      0.03389

> Thanks,
>
> Roger
> ***************************************************************
>
> This message is for the named person's use only. It may\...{{dropped: 
> 14}}
>
> ______________________________________________
> 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.
David Winsemius, MD
West Hartford, CT

have a look at ?formula and the examples

2011/9/8 Bos, Roger <roger.bos at rothschild.com>:
> I have a matrix called mat and y is the column number of my response and
> x is a vector of the column numbers of my terms. ?The variable name of y
> can change, so I don't want to hardcode it. ?I can find out the name as
> follows:
>
>> ? ? ? ? ? ?names(mat)[y]
> [1] "er12.l"
>
> Then I can run the regression by hard coding the variable name as
> follows:
>
>> ? ? ? ? ? ? mod <- try(rlm(er12.l ~ ., data=mat[zidx, c(y, x)]),
> silent=TRUE)
>
> But how would I do so without hard coding the name er12.l?
>
> I set up a reproducible example. ?In the following my regression formula
> is aa ~ bb + cc, or more simply aa ~ .
> How can I use the name of the first column without hard coding aa?
>
> dat <- data.frame(aa=runif(50), bb=runif(50), cc=runif(50))
> names(dat)[1]
> lm(aa ~ ., data=dat[1:10,])
>
> Thanks,
>
> Roger
> ***************************************************************
>
> This message is for the named person's use only. It
may\...{{dropped:14}}
>
> ______________________________________________
> 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.
>

The answer to my question was :

fmla <- as.formula(names(mat)[y] %+% " ~ .")
mod <- try(rlm(fmla, data=mat[zidx, c(y, x)]), silent=TRUE) 

Thanks for your help and the quick response.

Roger

-----Original Message-----
From: Jean-Christophe BOU?TT? [mailto:jcbouette at gmail.com] 
Sent: Thursday, September 08, 2011 9:22 AM
To: Bos, Roger
Cc: r-help at r-project.org
Subject: Re: [R] How to specify a variable name in the regression formula
without hard coding it

have a look at ?formula and the examples

2011/9/8 Bos, Roger <roger.bos at rothschild.com>:
> I have a matrix called mat and y is the column number of my response 
> and x is a vector of the column numbers of my terms.  The variable 
> name of y can change, so I don't want to hardcode it.  I can find out 
> the name as
> follows:
>
>>            names(mat)[y]
> [1] "er12.l"
>
> Then I can run the regression by hard coding the variable name as
> follows:
>
>>             mod <- try(rlm(er12.l ~ ., data=mat[zidx, c(y, x)]),
> silent=TRUE)
>
> But how would I do so without hard coding the name er12.l?
>
> I set up a reproducible example.  In the following my regression 
> formula is aa ~ bb + cc, or more simply aa ~ .
> How can I use the name of the first column without hard coding aa?
>
> dat <- data.frame(aa=runif(50), bb=runif(50), cc=runif(50)) 
> names(dat)[1] lm(aa ~ ., data=dat[1:10,])
>
> Thanks,
>
> Roger
> ***************************************************************
>
> This message is for the named person's use only. It ...{{dropped:26}}