reweightedQn: Reweighted Variant of the Robust Scale Estimator Qn
In robts: Robust Time Series Analysis

Description Usage Arguments Details Value Author(s) References See Also

Computes a reweighted variant of the robust scale estimator Qn.

1	reweightedQn(x, boundq = 0.999)

`x`	numeric vector of observations.
`boundq`	numeric value specifying the quantile of the chi square distribution for the reweighting step, see Details.

In a first step the ordinary Qn estimator is computed. In a second step the empirical standard deviation is computed using all observations whose squared residuals based median and Qn are smaller than the boundq quantile of a chi square distribution with one degree of freedom. This additional reweighting step improves the efficiency as compared to the Qn estimator, which was introduced by Rousseeuw and Croux (1993).

Numeric value with the estimated scale.

Alexander Dürre, Tobias Liboschik and Jonathan Rathjens

Rousseeuw, P.J., Croux, C. (1993): Alternatives to the median absolute deviation, Journal of the American Statistical Association, vol. 88, 1273–1283.

Qn for the non-reweighted Qn estimator.

robts documentation built on May 2, 2019, 4:55 p.m.