MultivariateStatistics`

MultivariateTrimmedMean

gives the mean of the bivariate data matrix after dropping a fraction f of the outermost vectors.

Details and Options

To use MultivariateTrimmedMean, you first need to load the Multivariate Statistics Package using Needs["MultivariateStatistics`"].
MultivariateTrimmedMean gives a robust estimate of the mean by excluding extreme values.
The outlying vectors are removed by repeatedly peeling off layers of convex hulls from the data until at least a fraction f have been removed.
MultivariateTrimmedMean interpolates between the means of the points remaining before and after the last layer is removed.
MultivariateTrimmedMean[matrix,0] is equivalent to Mean[matrix].
MultivariateTrimmedMean[matrix,f] approaches ConvexHullMedian[matrix] as f approaches 1.

Multivariate trimmed mean of bivariate data:

Top