Skip to contents

Biweight Midcovariance

Source: R/biweight_midcovariance.R
biweight_midcovariance.Rd

This function computes the biweight midcovariance, a robust measure of covariance between two numerical vectors. The biweight midcovariance is less sensitive to outliers than the traditional covariance.

Usage

biweight_midcovariance(x, y)

Arguments

x

A numeric vector.

y

A numeric vector of the same length as x.

Value

The biweight midcovariance between x and y.

References

  • Wilcox, R., (1997). Introduction to Robust Estimation and Hypothesis Testing. Academic Press

Author

Christian L. Goueguel

Examples

data(forageLIBS)
# iron and manganese contents (mg/kg), with a few iron-rich samples
ok <- !is.na(forageLIBS$Fe)
c(covariance = stats::cov(forageLIBS$Fe[ok], forageLIBS$Mn[ok]),
  biweight = biweight_midcovariance(forageLIBS$Fe[ok], forageLIBS$Mn[ok]))
#> covariance   biweight 
#>   1146.068    322.688