cov.shr: Yohai's Smoothed Hard Rejection Multivariate M-Estimator

In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Description

This implements the 'smoothed hard rejection' estimator. It is very similar to Rocke's translated biweight S-estimator in the cov.shr function, but has been found to work better in problems where the number of variables is greater than 15 or so (Maronna & Yohai, 2017). The algorithm is initialized with Billor, Hadi, and Velleman's (2000) BACON algorithm.

Usage

cov.shr(X, maxit = 50, tol = 1e-04)

Arguments

X	a data frame or matrix of numeric covariates
maxit	maximum number of iterations. defaults to 50.
tol	convergence tolerance

Value

a covRobust object containing the following elements:

• center: multivariate mean of cleaned data set after applying casewise weights.

• cov: covariance matrix of cleaned data set after applying casewise weights.

• dist: the mahalanobis distances used in calculating the weights.

• outliers: the indices of the outliers identified.

• weights: the weights for downweighting outliers.

References

Muler, N. & Yohai, V.J. (2002). Robust estimates for arch processes. Journal of Time Series Analysis, 23(3), 341–375. doi:10.1111/1467-9892.00268

Maronna, R.A. & Yohai, V.J. (2017) Robust and efficient estimation of multivariate scatter and location. Computational Statistics and Data Analysis, 109, 64–75. doi: 10.1016/j.csda.2016.11.006

Billor, N., Hadi, A. S., & Velleman , P. F. (2000). BACON: Blocked Adaptive Computationally-Efficient Outlier Nominators; Computational Statistics and Data Analysis, 34, 279–298. doi: 10.1016/S0167-9473(99)00101-2

abnormally-distributed/cvreg documentation built on May 3, 2020, 3:45 p.m.