Abstract
This paper considers algorithms for solving the linear robust regression problem by minimizing the Huber function. In the computational methods for this problem used so far, the scale estimate is adjusted separately. The new algorithm, based on Newton's method, treats both the scale and the location parameters as independent variables. The special form of the Hessian allows for an efficient updating scheme.
Similar content being viewed by others
References
D. I. Clark and M. R. Osborne,Finite algorithms for Huber's M-estimator, SIAM J. Scient. Stat. Comp. 7 (1986), 72–85.
J. J. Dongarra et al.,LINPACK Users Guide, SIAM Publications, Philadelphia, 1979.
R. Dutter,Numerical solution of robust regression problems: Computational aspects, a comparison, J. Statist. Comput. Simul. 5 (1977), 207–238.
H. Ekblom,Programs for the Huber estimator in linear models, Dept. of Math. 1985–5, Luleå University, Sweden.
F. R. Hampel, E. M. Ronchetti, P. J. Rousseeuw and W. A. Stahel,Robust Statistics: The Approach Based on Influence Functions, John Wiley, New York, 1986.
P. Huber,Robust Statistics, John Wiley, New York, 1981.
A. Marazzi,On the Numerical Solutions of Bounded Influence Regression Problems, in F. de Antonio, N. Laauro and A. Rizzi (eds.),COMPSTAT 86: Proceedings in Computational Statistics, Physica Verlag, Vienna, 1986.
Nag Reference Manual, Mark 10, NAG Central Office, Oxford, England, 1983.
J. M. Ortega and W. C. Reinboldt,Iterative Solution of Nonlinear Equations in Several Variables, Academic Press, London and New York, 1970.
A. Ruhe and P. Å. Wedin,Algorithms for separable nonlinear least-squares problems, SIAM Review 22 (1980), 318–337.
D. F. Shanno and K. H. Phua,Research on Algorithm 500: Minimization of Unconstrained Multivariate Functions, ACM Trans. Math. Software 6 (1980), 618–622.
D. F. Shanno and D. M. Rocke,Numerical methods for robust regression: Linear models, SIAM J. Scient. Stat. Comp. 7 (1986), 86–97.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Ekblom, H. A new algorithm for the huber estimator in linear models. BIT 28, 123–132 (1988). https://doi.org/10.1007/BF01934700
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01934700