movMean: Weighted Simple Moving Mean
In RcppMovStat: Fast Moving Statistics Calculation

Description Usage Arguments Details Value Functions Examples

View source: R/RcppExports.R

This function returns a simple moving average of the given vector. The weight argument is optional.

movMean(vec, n = 1L, ss = 1L, w = NULL, na_rm = FALSE, sizeD = FALSE,
  align = "left")

movMeanr(vec, n = 1L, ss = 1L, w = NULL, na_rm = FALSE, sizeD = FALSE)

`vec`	A numeric vector.
`n`	An integer: moving window size, with 1 as default
`ss`	An integer: step size, only calculating at points with an equal distance ss. Namely, there are ss-1 number between each two 'consecutive' points
`w`	An optional weight vector of length n. It will be automatically normalized (sum to 1).
`na_rm`	logical. Should missing values (including NaN) be removed?
`sizeD`	logical. Only applied when ss > 1, it decides whether to get a result of smaller size. If sizeD = T, align does not affect the output.
`align`	A string denotes how to align the moving average, three options: "left", "middle", "right"

Despite of Efficient computation, usually 5~6 times faster than the moving average function in package 'RcppRoll', it is able to handle potential missing values (NA or NaN) in the vec.
For instance, the output of the second example is NA, NA, 2.200000 3.714286 4.875000. The last number 5.5 is obtained by using re-normalized weight, namely omitting 0.2. The weight applied would be 0.5/(0.5+0.3) and 0.3/(0.5+0.3). Hence,

4.875 = 3 * 0.5/(0.5+0.3) + 8 * 0.3/(0.5+0.3)

This function returns a vector whose length is the same as that of vec or is ceiling((L - n + 1)/ss), (when sizeD = T), where L is the length of vec.

movMeanr: An function equivalent to movMean(..., align = "right")

1
2
3

movMean(c(1, 4, 3, NA, 8), 3, align = "right", na_rm = TRUE)
movMean(c(1, 4, 3, NA, 8), 3, w = c(0.5, 0.2, 0.3), na_rm = TRUE, align = "right")
movMean(c(1, 4, 3, NA, 8, 4, 5, 9, 6, 0), n = 3, ss = 4, na_rm = TRUE, align = "right")