Passing list of values as parameters to a function in R

Question

I have two lists

   a=expand.grid(mtry=c(8))
   b=expand.grid(usekernel=c(8), fL=c(9))
   c=expand.grid(usekernel=c(8), fL=c(9))
   list_1<- list(a,b,c)
   list_2<- list("rf","nb","nb2")

And I have a function as follows

  func(TrainData, TrainClasses,method,tune)
  if method is "rf" then tune=a, if "nb" then tune=b, nb2=c

Instead of writing the method and tune every time, I want to pass the lists(list_1 and list_2) and invoke the "func" function for every method. I tried for loop, but its not giving the desired results.

Reproducible code

library(caret)
data(iris)
TrainData <- iris[,1:4]
TrainClasses <- iris[,5]
method<- c("rf")
tune<-expand.grid(mtry=c(8))

func<- function(TrainData, TrainClasses,method,tune )
{
 k <- train(TrainData, TrainClasses,
               method = method,
               preProcess = c("center", "scale"),
               tuneLength = 10,
               tuneGrid = tune,
               trControl = trainControl(method = "cv"))
 k
}

func(TrainData, TrainClasses,method, tune)

So, in the above function, I'm passing one method and its tune, but now I want to pass the list of methods one after the other with their respective tune instead of giving method and its tune every time, I want to execute all the 3 methods at once. Like I said before, my list_2 contains methods and list_1 contains respective tunes.

My code:

 for(i in 1:3){
     tune_e=c(list(TrainData=TrainData,TrainClasses=TrainClasses),list_2[i],list_1[i])
     do.call('func',tune_e)
 }

Please create a [reproducible example](http://stackoverflow.com/a/5963610/1412059) to facilitate good answers. — Roland, Jul 17 '15 at 08:06
"I tried for loop" - please add your code even if it doesn't work, it will help us to see the logic. — zx8754, Jul 17 '15 at 08:08
@Roland I edited my post with reproducible code and scenario.. please check — Knight, Jul 17 '15 at 08:39
@Roland I'm sry, I added the library to the post, it comes from caret package — Knight, Jul 17 '15 at 09:15

CJB · Accepted Answer · 2015-07-17T09:26:21.520

2

Excuse me if I haven't quite grasped what you want. Would this work?

list_1 <- c(a,b,c); list_2 <- c("rf","nb","nb2")
for(i in 1:3){
    func(TrainData, TrainClasses, list_2[i], list_1[i])
}

The problem is with the lists. If they are simply vectors, then it should work. Do they need to be lists for another reason? If so:

list_1 <- list(a,b,c); list_2 <- list("rf","nb","nb2")
for(i in 1:3){
    func(TrainData, TrainClasses, list_2[[1]][i], list_1[[1]][i])
}

The double square brackets are important.

edited Jul 17 '15 at 09:26

answered Jul 17 '15 at 08:57

CJB

1,759
17
26

Also, why do you need to wrap the `train` function, since that is the only thing that is executed within `func`? – CJB Jul 17 '15 at 08:59
I updated my code in my post, I'm trying to add some more parameters and methods and post process to it if it works out – Knight Jul 17 '15 at 09:17
I tried you code its not giving the results as well – Knight Jul 17 '15 at 09:20
error : some required components are missing: library, type, parameters, grid, fit, predict, prob – Knight Jul 17 '15 at 09:22
I've edited now - one problem was the list elements. If you specify them as vectors, it is easier. Alternatively use `list_1[[1]][i]` to extract the elements of the list. – CJB Jul 17 '15 at 09:27
its not working, thing is here the method(list_2) can consider one variable at a time( for example list_2= rf, then if list_1=a(mtry) , but when we are considering vectors, it clubs three values like mtry, usekernal, fL(3 variables) – Knight Jul 17 '15 at 09:41
but really thanks for suggestions, I'll think about it :) – Knight Jul 17 '15 at 09:42
Then surely the list indexing system (`list_1[[1]][i]`) should work? – CJB Jul 17 '15 at 09:49
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/83530/discussion-between-maddy-and-christopher-barry). – Knight Jul 17 '15 at 10:02

Passing list of values as parameters to a function in R

1 Answers1