Distributing a computationally intensive estimator: the case of exact LMS regression

MaRDI QIDQ1965947zbMATH OpenFDO

Authors Douglas M. Hawkins, Jeffrey S. Simonoff, Arnold J. Stromberg

Publication date 2 March 2000

Published in Computational Statistics (Search for Journal in Brave)

zbMATH Keywords

distributed architectureleast median of squaresparallel architecture

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Linear regression; mixed models (62J05) Computer aspects of numerical algorithms (65Y99)

Review text

Consider the usual linear regression model \(y_{i}=x'_{i}\beta+\varepsilon_{i}\), \(i=1,\dots,n\), where \(\beta\) is an unknown vector-valued parameter and \(\varepsilon_{i}\) are unknown errors. The least median of squares estimator \(\hat \beta_{LMS}\) is defined as \[ \underset\beta{\text{Arg min}} \underset{1\leq i\leq n}{\text{Median}} (y_{i}-x'_{i}\beta)^{2}. \] A set of Fortran subroutines for determining the exact \(\hat\beta_{LMS}\) is presented. Details are given for distributing the computations. Implementation of this algorithm for a network of computers is discussed. The proposed distribution techniques can be directly applied to any estimator whose calculation is based on repeated subsampling from the data.

Recommendations

• Robust and scalable distributed recursive least squares
• On the efficient computation of robust regression estimators
• Distributed optimization and statistical learning for large-scale penalized expectile regression
• Communication-efficient distributed estimator for generalized linear models with a diverging number of covariates
• Almost exact distributions of estimators I-low dimensional nonlinear regression
• Distributionally robust \(L_1\)-estimation in multiple linear regression
• Efficient estimation of the error distribution in a varying coefficient regression model

Cited in

(8)

• Regression-free and robust estimation of scale for bivariate data
• Parallel algorithms for least median of squares regression
• Reweighting approximate GM estimators: Asymptotics and residual-based graphics
• BACON: blocked adaptive computationally efficient outlier nominators.
• Least trimmed squares regression, least median squares regression, and mathematical program\-ming
• Formulas for the exact LMS and LQS estimator
• Analyzing the number of samples required for an approximate Monte-Carlo LMS line estimator
• Convergence of the feasible solution algorithm for least median of squares regression

This page was built for publication: Distributing a computationally intensive estimator: the case of exact LMS regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1965947)