cleanOutliers: Clean Returns Outliers Effectively

In PCRA: Companion to Portfolio Construction and Risk Analysis

Clean Returns Outliers Effectively

Description

Outliers are "cleaned" by shrinking or rejecting data whose distance from the median (med) is larger in absolute value than a specified value k multiplied by the median absolute deviation from the median (mad). Outlier shrinkage results in the data value being set equal to the nearest of med-k*mad and med+k*mad. Rejected data is assigned an NA. Shrinkage is the default.

Usage

cleanOutliers(x, k = 3, shrink = TRUE)

Arguments

x	A numeric vector
k	A numeric value, which multiplies the mad. Smaller values of k result in greater fractions of data which is either shrunk of rejected, and larger values of k result in smaller fractions of the data that are shrunk or rejected.
shrink	A logical variable whose default is TRUE.

Value

an outlier cleaned numeric object

Examples

args(cleanOutliers)

PCRA documentation built on July 15, 2026, 9:06 a.m.