R help - Oct 2010 - Help with function writing

Hello all
I have what seems like a simple question but have not been able to find an
answer on the forum. I'm trying to define a function which involves
regression models and a large number of covariates. 

I would like the function to accept any number of covariates and, ideally, I
would like to be able to enter the covariates in a group (e.g. as a list)
rather than individually. Is there any way of doing this? 

Example:

#define function involving regression model with several covariates 
custom <- function(outcome, exposure, covar1, covar2, covar3){
  model <- lm(outcome ~ exposure + covar1 +  covar2 + covar3)
  expected <- predict(model)
  summary(expected)
}

library(MASS)
attach(birthwt)

custom(bwt, lwt, low, age, race) #Works when 3 covariates are specified

custom(bwt,lwt,low,age) # Does not work with < or > 3 covariates

varlist <- list(low,age,race)
custom(bwt,lwt, varlist) #Does not work if covariates are included as a list



Thanks very much for your help

Tim

--
Tim Elwell-Sutton
University of Hong Kong

Hello Tim,

This function will do it where the covariates are provided as separate
arguments. It would be easy to modify this to handle a list too.

function(outcome, ...) {
  arg.names <- as.character(match.call())[-1]
  nargs <- length(arg.names)
  f <- as.formula(paste(arg.names[1], "~",
paste(arg.names[2:nargs],
collapse="+")))

  model <- lm(f)

  # rest of your code here
}

Hope that helps you get started.

Michael


On 12 October 2010 14:35, Tim Elwell-Sutton <tesutton at hku.hk>
wrote:
> Hello all
> I have what seems like a simple question but have not been able to find an
> answer on the forum. I'm trying to define a function which involves
> regression models and a large number of covariates.
>
> I would like the function to accept any number of covariates and, ideally,
I
> would like to be able to enter the covariates in a group (e.g. as a list)
> rather than individually. Is there any way of doing this?
>
> Example:
>
> #define function involving regression model with several covariates
> custom <- function(outcome, exposure, covar1, covar2, covar3){
> ?model <- lm(outcome ~ exposure + covar1 + ?covar2 + covar3)
> ?expected <- predict(model)
> ?summary(expected)
> }
>
> library(MASS)
> attach(birthwt)
>
> custom(bwt, lwt, low, age, race) #Works when 3 covariates are specified
>
> custom(bwt,lwt,low,age) # Does not work with < or > 3 covariates
>
> varlist <- list(low,age,race)
> custom(bwt,lwt, varlist) #Does not work if covariates are included as a
list
>
>
>
> Thanks very much for your help
>
> Tim
>
> --
> Tim Elwell-Sutton
> University of Hong Kong
>
> ______________________________________________
> 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.
>