Skip to contents

Flag outliers with robust Z-scores

Source: R/flag_z.R
flag_z.Rd

Identifies potential outliers in a numeric vector based on a moving window robust Z-score approach. The robust Z-score is computed using a biweight scale estimate centered on the median.

Usage

flag_z(x, width = 5, threshold = 3, return_z = FALSE)

Arguments

x

Numeric vector to check for outliers.

width

Odd integer giving the moving-window width. Defaults to 5.

threshold

Numeric threshold for the absolute Z-score above which a value is flagged. Defaults to 3.

return_z

Logical. If TRUE, return both Z-scores and flags. If FALSE, return only flags. Defaults to FALSE.

Value

If return_z = TRUE, a list with:

z

A numeric vector of robust Z-scores (with NA where not computable).

flag

A character vector of the same length as x, with "Z" where an outlier is detected and NA otherwise.

If return_z = FALSE, only the flag vector is returned.

Details

For each value in x, a window of length width centered on that value is extracted. The function:

  • Computes the median of the window.

  • Calculates residuals from the median.

  • Estimates a robust scale using Tukey’s biweight estimator based on MAD.

  • Computes a Z-score as \((x_i - \text{median}) / \text{scale}\).

If the absolute value of the Z-score exceeds threshold, the value is flagged with "Z".

NA values are ignored in the window statistics but retained in output positions.

See also

stats::mad(), stats::median()

Examples

x <- c(1, 2, 1.5, 1.2, 100, 1.1, 1.3, 1.4)
flag_z(x)
#> [1] NA  NA  NA  NA  "Z" NA  NA  NA 
flag_z(x, return_z = TRUE)
#> $z
#> [1]  -0.8538437   2.8070970   0.0000000  -0.5964796 459.3115264  -1.2804234
#> [7]  -1.0015789   0.6289591
#> 
#> $flag
#> [1] NA  NA  NA  NA  "Z" NA  NA  NA 
#>