A Monte Carlo Study of Two Robust Alternatives of Least Square Regression Estimation

We give some Monte Carlo results on the performance of two robust alternatives to least squares regression estimation - least absolute residuals and the one-step "sine" estimator. We show how to scale the residuals for the sine estimator to achieve constant efficiency at the Gaussian across various choices of X-matrix and give some results for the contaminated Gaussian distribution.

MoreLess

Year of publication:	1974-09
Authors:	Hill, Richard W. ; Holland, Paul W.
Institutions:	National Bureau of Economic Research (NBER)

More details

Extent:	application/pdf
Series:	NBER Working Papers.
Type of publication:	Book / Working Paper
Notes:	published as Hill, Richard W. and Paul W. Holland. "Two Robust Alternatives To Least-Squared Regression," Journal of the American Statistical Association, 1977, v72(360), 828-833. Number 0058
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10005830590