Selecting columns in a data.frame to implement in a model

Question

Is there a way to run a model (for simplicity, a linear model) using specified columns of a data.frame?

For example, I would like to be able to do something like this:

set.seed(1)

ET = runif(10, 1,20)

x1 = runif(10, 1,20)

x2 = runif(10, 1,30)

x3 = runif(10, 1,40)



Xdf = data.frame(ET = ET, X1 = x1, X2 =x2, X3 = x3)



lm(ET~Xdf[,c(2,3)], data = Xdf)

Where the linear model would be equal to

lm(ET~X1 +X2, data = Xdf)

I have tried with a matrix - but it won't work in this case as I will eventually be adding spatial correlation based upon values stored in the data.frame that need to be specified by the data = data.frame call.As well as having certain names.frame. As well, I need to be able to choose certain columns in the data because this will be looping through multiple models using different predictors.

Any help would be greatly appreciated. Thanks!

Answer 1

Here's a (rather ugly) way to make it work.

I use as.formula and the paste function to make the formula before calling lm.

I'm sure there are better ways to do this, but this is what came to mind.

# ET ~ X1 + X2
f1 <- as.formula(paste("ET~", paste(names(Xdf)[c(2,3)], 
                                        collapse="+")))
lm(f1, data=Xdf)

You can also specify the individual columns, though it might be more work

lm(ET ~ Xdf[,2] + Xdf[,3], data=Xdf)

Selecting columns in a data.frame to implement in a model

Sign up