RejectOutlier: Finds the "worst" outlier in a multivariate time series

In R-Finance/Meucci: Collection of functionality ported from the MATLAB code of Attilio Meucci.

Description

Finds the "worst" outlier in a multivariate time-series We aim at finding the the observation x_t such that if we remove it from the set x_1 ... x_t the determinant of the resulting sample covariance is reduced the most This means that by dropping that observation the location-dispersion ellipsoid defined by the sample mean and covariance shrinks the most See Sec. 4.6.1 of "Risk and Asset Allocation" - Springer (2005), by A. Meucci for the theory and the routine implemented below

Usage

  RejectOutlier(sample)

Arguments

sample

a vector containing the input dataset with outliers

Value

rejected a numeric indicating which observation in the index to reject

Author(s)

Ram Ahluwalia [e-mail]

References

http://www.symmys.com See Meucci script for "RejectOutlier.m"

R-Finance/Meucci documentation built on May 8, 2019, 3:52 a.m.