Construction of Repeated loop in R.

#1

At first step, I want to generate a series (more specifically; GARCH (1, 1)), and the estimated it with four different estimation methods (let’s say, M1, M2, M3, M4). In addition, I want to **repeat this process over 6000 times with saving results of each repetition.

I have done it one by one, but unable to construct a loop.

0 Likes

#2

Hi Farrukh,

We don't really have enough info to help you out. Could you ask this with a minimal REPR oducible EX ample (reprex)? A reprex makes it much easier for others to understand your issue and figure out how to help.

1 Like

#3

Let's say you've a function estimate_GARCH_parameters_by_four_methods, which returns the estimates in a vector.

Then I suppose it'll do what you want:

results <- list()
counter <- 1
while(counter < 6000) {
    results[[counter]] <- estimate_GARCH_parameters_by_four_methods(...)
    counter <- (counter + 1)
}

Here, ... are the arguments of your function.

Please provide more information (preferably in form of a reprex, as Jim suggested) in order to get more specific help.

0 Likes

#4

"Growing" a vector can dramatically slow down code. How much depends on the code, but it's better to avoid it all the time. If you know the "sizes" of the result, create an object with those sizes and fill it in as you go.

@Yarnabrina's code can be rewritten as:

results <- vector("list", 6000)
for (counter in seq_along(results)) {
    results[[counter]] <- estimate_GARCH_parameters_by_four_methods(...)
}

A for loop is good for "complex" logic inside the loop. If the steps are just function calls, lapply() would be "R idiomatic." But, here, the code depends on the logic. So more details are needed.

0 Likes

#5

I didn't quite understand. Can you explain why these two won't work?

result <- lapply(X = 1:6000,
                 FUN = function(t) estimate_GARCH_parameters_by_four_methods(...))

or,

result <- sapply(X = 1:6000,
                 FUN = function(t) estimate_GARCH_parameters_by_four_methods(...))
0 Likes

#6

It depends on what's actually being done. If there really is a function like estimate_GARCH_parameters_by_four_methods(...), then lapply() is a great solution.

But if there isn't, then lapply() often won't simplify code. For example:

results <- vector("list", 6000)

for (ii in seq_along(results)) {
  # A whole bunch of prep logic
  m1 <- estimate_GARCH_parameters_m1(...)
  m2 <- estimate_GARCH_parameters_m2(...)
  m3 <- estimate_GARCH_parameters_m3(...)
  m4 <- estimate_GARCH_parameters_m4(...)
  results[[ii]] <- list(m1, m2, m3, m4)
}

Writing a wrapper function and feeding it to lapply() won't clarify this code.

estimate <- function(ii) {
  # A whole bunch of prep logic
  m1 <- estimate_GARCH_parameters_m1(...)
  m2 <- estimate_GARCH_parameters_m2(...)
  m3 <- estimate_GARCH_parameters_m3(...)
  m4 <- estimate_GARCH_parameters_m4(...)
  list(m1, m2, m3, m4)
}

results <- lapply(1:6000, estimate)

The for block is obviously just an iterative loop, where the 1:6000 vector might be irrelevant. The lapply() code makes the 1:6000 vector seem important.

0 Likes

closed #7

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.

0 Likes