I'm a beginner and I'm trying to write a function that fits a model (univariate) one variable at a time. I want to able to enter all the dependent variables when calling the function and then make the function to fit one variable at a time. Something like:
f <-function(y, x1,x2,x3) {(as.formula(print()))...}
I tried to make a list of x1, x2, x3 inside the function but it didn't work.
Maybe someone here could help me with this or point me in the right direction.
Here's something that might get you going. You can put your x variables inside a list, and then fit once for each element inside the list.
fitVars <- function(y, xList) {
lapply(xList, function(x) lm(y ~ x))
}
y <- 1:10
set.seed(10)
xVals <- list(rnorm(10), rnorm(10), rnorm(10))
fitVars(y, xVals)
This returns one fit result for each element of xVals:
[[1]]
Call:
lm(formula = y ~ x)
Coefficients:
(Intercept) x
4.984 -1.051
[[2]]
Call:
lm(formula = y ~ x)
Coefficients:
(Intercept) x
5.986 -1.315
[[3]]
Call:
lm(formula = y ~ x)
Coefficients:
(Intercept) x
7.584 2.282
Another option is to use the ... holder to use an arbitrary number of arguments:
fitVars <- function(y, ...) {
xList <- list(...)
lapply(xList, function(x) lm(y ~ x))
}
set.seed(10)
fitVars(y, rnorm(10), rnorm(10), rnorm(10))
This gives the same result as above.
