Abstract
Several methods have been suggested for obtaining the estimates of the solution vector for the robust linear regression model Ax=b+ε using an iteratively reweighted criterion based on weighting functions, which tend to diminish the influence of outliers. We consider a combination of Newton method (or iteratively reweighted least squares method) with a Krylov subspace method. We show that we can avoid preconditioning with A T A preconditioner by merely transforming the sequence of linear systems to be solved. Appropriate sequences of linear systems for sparse and dense matrices are given. By employing efficient sparse QR factorization methods, we show that it is possible to solve efficiently large sparse linear systems with an unpreconditioned Krylov subspace method.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Baryamureeba, V.: The impact of equal-weighting of both low-confidence and highconfidence observations on robust linear regression computation. BIT 41, 847–855 (2001)
Baryamureeba, V.: Solution of large-scale weighted least squares problems. Numerical Linear Algebra with Applications 9, 93–106 (2002)
Baryamureeba, V.: Solution of robust linear regression problems by preconditioned conjugate gradient type methods. Technical Report No. 187, Department of Informatics, University of Bergen, 5020 Bergen, Norway, March 22 (2000)
Baryamureeba, V., Steihaug, T.: On the properties of preconditioners for robust linear regression. Technical Report No. 184, Department of Informatics, University of Bergen, Norway. Journal of Opt. Theory and Appl. (March 2000) (to appear)
Dembo, R.S., Eisenstat, S.C., Steihaug, T.: Inexact Newton methods. SIAM Journal of Numerical Analysis 19, 400–407 (1982)
Fair, R.C.: On the robust estimation of econometric models. Ann. Econ. Social Measurement. 3, 667–678 (1974)
Hinich, M.J., Talwar, P.P.: A simple method for robust regression. J. Amer. Statist. Assoc. 70, 113–119 (1975)
Huber, P.J.: Robust estimation of a location parameter. Annals Math. Statist. 35, 73–101 (1964)
Huber, P.J.: Robust statistics. J. Wiley, New York (1981)
O’Leary, D.P.: Robust regression computation using iteratively reweighted least squares. SIAM Journal of Matrix Analysis and Applications 11, 466–480 (1990)
Scales, J.A., Gersztenkorn, A., Treitel, S.: Fast l p solution of large, sparse, linear systems: Application to seismic travel time tomography. Journal of Computation Physics 75, 314–333 (1988)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Baryamureeba, V. (2004). Solution of Robust Linear Regression Problems by Krylov Subspace Methods. In: Lirkov, I., Margenov, S., Waśniewski, J., Yalamov, P. (eds) Large-Scale Scientific Computing. LSSC 2003. Lecture Notes in Computer Science, vol 2907. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-24588-9_6
Download citation
DOI: https://doi.org/10.1007/978-3-540-24588-9_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-21090-0
Online ISBN: 978-3-540-24588-9
eBook Packages: Springer Book Archive