迭代地将变量添加到 R 中的 lm() function？

Question

Simple question, but I'm finding myself boggled.

I'm looking to make a loop that will continuously add variables to the IV of an lm() function. I would test the results of the LM until a condition is met. I'm just having trouble finding a way to dynamically adding variables to the IV part of the regression, one at a time.

The 1st iteration would look like:

lm(Y ~ X, data = data)

The second iteration like:

lm(Y ~ X + X2, data = data)

The third iteration like:

lm(Y ~ X + X2 + X3, data = data)

And so on...

If any of you could point me in the right direction, I'd appreciate it very much.

Thanks!

Answer 1

An alternative way is to use Y ~. as the formula and provide the subset of data as required. Here, . means "all columns not otherwise in the formula" (see ?formula ). Using mtcars as an example:

Y <- 'mpg'
Xs <- names(mtcars)[-1]

fits <- lapply(seq_along(Xs), function(x){
    lm(paste(Y, '~ .'), data = mtcars[, c(Y, Xs[1:x])])
})

Answer 2

We can use reformulate to create the formula after passing the independent variables as a list

out <- lapply(list("X", c("X, "X2"), c("X", "X2", "X3")), 
      function(x) lm(reformulate(x, response = "Y"), data = data))

Or make it automated

Xs <- setdiff(names(data), "Y")
ind <- sequence(seq_along(Xs))
lapply(split(Xs[ind], cumsum(ind == 1)), function(x) 
          lm(reformulate(x, response = "Y"), data = data))