How to loop a formula in a regression model in R

Question

Suppose I have a data such like this:

n=100
df<-data.frame(y=rnorm(n,2,3),
               x1=rbinom(n,1,0.3),
               x2=rbinom(n,10, 0.5), 
               x3=rnorm(n, 50, 20),
               x4=rnorm(n, 3, 2))

and I have a basic model y=x1+x2:

mod0<-as.formula(y~x1+x2)
lm0<-lm(mod0, data=df)

what I want is to update the mod0 with an interaction term between x1 and each of x3 and x4 such that mod1<-y~x1+x2+x1*x3 and mod2<-y~x1+x2+x1*x4. Since I have a bunch of variables need to loop through, I am wondering what it may be a best way for that.

See http://stackoverflow.com/questions/30264666/looping-over-list-of-items/30264912#30264912 — Sam Firke, May 20 '15 at 16:53

score 5 · Accepted Answer · answered May 20 '15 at 16:03

5

Rather than assigning loose variables in the global environment, it would be better to assemble the formulae in a list. This can be done with lapply(), and leveraging the useful trick that allows you to create a formula from a character string:

lapply(names(df)[-(1:3)],function(var) formula(paste0('y~x1+x2+x1*',var),env=globalenv()));
## [[1]]
## y ~ x1 + x2 + x1 * x3
##
## [[2]]
## y ~ x1 + x2 + x1 * x4

answered May 20 '15 at 16:03

bgoldst

34,190
6
38
64

Interesting answer, does your solution apply to [this similar question](https://stackoverflow.com/questions/58756403/combinations-of-variables-that-produce-the-smallest-quantities-in-an-r-function)? – rnorouzian Nov 08 '19 at 01:30

How to loop a formula in a regression model in R

1 Answers1