创建一个循环来计算不同滚动平均值长度的矢量的滚动平均值

Question

I am trying to create a "for loop" setup that is going calculate different rolling means of a return series, where I use rolling means ranging from the last 2 observations to the last 16 observations. kϵ[2,16]. I've been trying to use a function like this, where the "rollmean" is a function from zoo. This produces the warning "Warning message: In roll[i] <- rollmean(x, i) : number of items to replace is not a multiple of replacement length" Can someone please help me?

rollk <- function(x, kfrom= 2, kto=16){
roll <- as.list(kto-kfrom+1)
for (i in kfrom:kto){
roll[i]<- rollmean(x, i)
return(roll)
}}

Answer 1

I suppose you want

# library(zoo)
rollk <- function(x, kfrom = 2, kto = 16){
  roll <- list()
  ft <- kfrom:kto
  for (i in seq_along(ft)){
    roll[[i]]<- rollmean(x, ft[i])   
  }
  return(roll)
}

There are several problems in your function:

• You need [[ to access a single list element, not [ .
• You want a list of length length(krom:kto) . Now, i starts at 1 , not at kfrom .
• Now, roll is returned after the for loop. Hence, the function returns a single list containing all values.

---

A shorter equivalent of the function above:

rollk2 <- function(x, kfrom = 2, kto = 16)
  lapply(seq(kfrom, kto), function(i) na.omit(filter(x, 1 / rep(i, i))))

It does not require loading additional packages.

Answer 2

Try this:

library(zoo)
lapply(2:16, rollmean, x = x)