weights.cov_Huber: Robustness weights of Huber M-estimation of location and...

In robmed: (Robust) Mediation Analysis

Robustness weights of Huber M-estimation of location and scatter

Description

Extract (relative) robustness weights of a Huber M-estimate of location and scatter.

Usage

## S3 method for class 'cov_Huber'
weights(object, type = c("consistent", "relative"), ...)

Arguments

object	an object inheriting from class "cov_Huber" containing Huber M-estimates of location and scatter.
type	a character string specifying the type of robustness weights to be extracted. Possible values are "consistent" and "relative". The former can be used for a robust transformation of the data such that the covariance matrix of the transformed data is Fisher consistent. Observations that are not downweighted in general receive a weight larger than 1. The latter are useful for interpretation, as observations that are not downweighted receive a relative weight of 1.
...	additional arguments are currently ignored.

Value

A numeric vetor containing the requested robustness weights.

Author(s)

Andreas Alfons

References

Zu, J. and Yuan, K.-H. (2010) Local Influence and Robust Procedures for Mediation Analysis. Multivariate Behavioral Research, 45(1), 1–44. doi:10.1080/00273170903504695.

See Also

cov_Huber()

Examples

data("BSG2014")

# define variables
x <- "ValueDiversity"
y <- "TeamCommitment"
m <- "TaskConflict"

# compute Huber M-estimator
S <- cov_Huber(BSG2014[, c(x, y, m)])
weights(S, type = "relative")

robmed documentation built on July 9, 2023, 6:29 p.m.