R help - Mar 2010 - looping through predictors

Dear R-ers,
I have a data frame data with predictors x1 through x5 and the
response variable y.
I am running a simple regression:

reg<-lm(y~x1, data=data)

I would like to loop through all predictors. Something like:
predictors<-c("x1","x2",... "x10)
for(i in predictors){
  reg<-lm(y~i)
 etc.
}

But it's not working. I am getting an error:
Error in model.frame.default(formula = Y ~ x1 + x2 + x3 + i, data = sample,  :
  variable lengths differ (found for 'i')

How can I make it take predictor names in the lm formula?
Thank you!

-- 
Dimitri Liakhovitski
Ninah.com
Dimitri.Liakhovitski at ninah.com

Dimitri -
    Without commenting on the wiseness of such an approach,
here's one way to do what you want:

regs = lapply(predictors,function(var)lm(data$y~data[,var]))
names(regs) = predictors

Now

regs[['x1']] holds the lm output from the regression of y on x1,
regs[['x2']] holds the lm output from the regression of y on x2, etc.

Suppose you wanted to know what the slopes for each regressor were.
First, find what you want for one:

coef(regs[['x1']])[2]

Next, write a function to extract this information:

getslope = function(reg)coef(reg)[2]

Now use sapply to get all the slopes of the individual regressions:

sapply(regs,getslope)

Hope this helps.

 					- Phil Spector
 					 Statistical Computing Facility
 					 Department of Statistics
 					 UC Berkeley
 					 spector at stat.berkeley.edu




On Tue, 9 Mar 2010, Dimitri Liakhovitski wrote:
> Dear R-ers,
> I have a data frame data with predictors x1 through x5 and the
> response variable y.
> I am running a simple regression:
>
> reg<-lm(y~x1, data=data)
>
> I would like to loop through all predictors. Something like:
> predictors<-c("x1","x2",... "x10)
> for(i in predictors){
>  reg<-lm(y~i)
> etc.
> }
>
> But it's not working. I am getting an error:
> Error in model.frame.default(formula = Y ~ x1 + x2 + x3 + i, data = sample,
:
>  variable lengths differ (found for 'i')
>
> How can I make it take predictor names in the lm formula?
> Thank you!
>
> -- 
> Dimitri Liakhovitski
> Ninah.com
> Dimitri.Liakhovitski at ninah.com
>
> ______________________________________________
> 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.
>

I already answered a very similar question earlier today. It was in  
the context of a question about a different regression function, but  
the formula methods are generic. Look for "Question on passing in  
parameter to Cox hazard".

(That does not mean I approve of such fishing expeditions.)

-- 
David.

On Mar 9, 2010, at 6:26 PM, Dimitri Liakhovitski wrote:
> Dear R-ers,
> I have a data frame data with predictors x1 through x5 and the
> response variable y.
> I am running a simple regression:
>
> reg<-lm(y~x1, data=data)
>
> I would like to loop through all predictors. Something like:
> predictors<-c("x1","x2",... "x10)
> for(i in predictors){
>  reg<-lm(y~i)
> etc.
> }
>
> But it's not working. I am getting an error:
> Error in model.frame.default(formula = Y ~ x1 + x2 + x3 + i, data =  
> sample,  :
>  variable lengths differ (found for 'i')
>
> How can I make it take predictor names in the lm formula?
> Thank you!
>
> -- 
> Dimitri Liakhovitski
> Ninah.com
> Dimitri.Liakhovitski at ninah.com
>

David Winsemius, MD
West Hartford, CT