rolling loop (functionals)

[mcooganj]

Hi,

I have been working through your very good advanced R book, but think there may be a bug in the rollmean function in the functionals chapter.

I may be wrong, but i'm fairly sure that a centred moving average would be:

rollmean <- function(x, n){
    out <- rep(NA, length(x))
    offset <- trunc(n/2)
    for (i in (offset + 1):(length(x) - n + offset + 1)) {
        out[i] <- mean(x[(i - offset):(i + offset)])
    }
    out
}

The edits are changing upper edge of the range to length(x) - n + offset + 1, and changing the processing range to (i - offset):(i + offset) in the mean step.

Happy to be corrected.

NB: I assign the copyright of this contribution to Hadley Wickham.

mj

[hadley]

Makes sense - thanks!