BIT Numerical Mathematics

, Volume 31, Issue 4, pp 697–710 | Cite as

On the solution of the errors in variables problem using thel1 norm

  • G. A. Watson
  • K. F. C. Yiu
Part II Numerical Mathematics


A fundamental problem in data analysis is that of fitting a given model to observed data. It is commonly assumed that only the dependent variable values are in error, and the least squares criterion is often used to fit the model. When significant errors occur in all the variables, then an alternative approach which is frequently suggested for this errors in variables problem is to minimize the sum of squared orthogonal distances between each data point and the curve described by the model equation. It has long been recognized that the use of least squares is not always satisfactory, and thel1 criterion is often superior when estimating the true form of data which contain some very inaccurate observations. In this paper the measure of goodness of fit is taken to be thel1 norm of the errors. A Levenberg-Marquardt method is proposed, and the main objective is to take full advantage of the structure of the subproblems so that they can be solved efficiently.

AMS subject classification

65D10 65K05 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Barrodale, I. and Roberts, F. D. K.:An improved algorithm for discrete l 1 linear approximation, SIAM J Num Anal 10, (1973) 839–848.Google Scholar
  2. 2.
    Barrodale, I. and Roberts, F. D. K.:An efficient algorithm for discrete l 1 linear approximation with linear constraints, SIAM J Num Anal 15, (1978) 603–611.Google Scholar
  3. 3.
    Bartels, R. H. and Conn, A. R.:An approach to nonlinear l 1 data fitting, inNumerical Analysis Cocoyoc 1981, Lecture Notes in Mathematics 909, ed Hennart, J. P., Springer Verlag, Berlin, (1982) 48–58.Google Scholar
  4. 4.
    Boggs, P. T., Byrd, R. H. and Schnabel, R. B.:Stable and efficient algorithm for nonlinear orthogonal distance regression, SIAM J Sci Stat Comp 8, (1987) 1052–1078.Google Scholar
  5. 5.
    Boggs, P. T., Byrd, R. H., Donaldson, J. R. and Schnabel, R. B.:ODRPACK, Software for weighted orthogonal distance regression, ACM Trans Math Software 15, (1989) 348–364.Google Scholar
  6. 6.
    Boggs, P. T., Donaldson, J. R., Schnabel, R. B. and Spiegelman, C. H.:A computational examination of orthogonal distance regression, J Econometrics 38, (1988) 169–201.Google Scholar
  7. 7.
    Ekblom, H. and Madsen, K.:Algorithms for non-linear Huber estimation, Technical University of Denmark, Institute for Numerical Analysis, Report NI 87-06 (1987).Google Scholar
  8. 8.
    Fletcher, R.:Numerical experiments with an exact l 1 penalty function method, in Nonlinear Programming 4, eds. Mangasarian, O. L., Meyer, R. R. and Robinson, S. M.; Academic Press, New York, (1981) 99–129.Google Scholar
  9. 9.
    Fletcher, R.:An l 1 penalty method for nonlinear constraints, inNumerical Optimization 1984, eds Boggs, P. T., Byrd R. H. and Schnabel, R. B., SIAM Publications, Philadelphia, (1985) 26–40.Google Scholar
  10. 10.
    Fletcher, R.:Practical Methods of Optimization, 2nd edition, Wiley, Chichester (1987).Google Scholar
  11. 11.
    Hald, J. and Madsen, K.:Combined LP and quasi-Newton methods for nonlinear l 1 optimization, SIAM J Num Anal 22, (185) 68–80.Google Scholar
  12. 12.
    Mclean, R. A. and Watson, G. A.:Numerical methods for nonlinear discrete l 1 approximation problems, inNumerical Methods of Approximation Theory, eds Collatz, L., Meinardus, G. and Werner, H., ISNM 52, Birkhäuser Verlag, Berlin, (1980) 169–183.Google Scholar
  13. 13.
    Murray, W. and Overton, M. L.:A projected Lagrangian algorithm for nonlinear l, SIAM J Sci Stat Comp 2, (1981) 207–224.Google Scholar
  14. 14.
    Powell, M. J. D.:General algorithms for discrete nonlinear approximation calculations, inApproximation Theory IV, eds Chui, C. K., Schumaker, L. L. and Ward, J. D., Academic Press, New York, (1983) 187–218.Google Scholar
  15. 15.
    Schwetlick, H. and Tiller, V.:Numerical methods for estimating parameters in nonlinear models with errors in the variables, Technometrics 27, (1985) 17–24.Google Scholar
  16. 16.
    Watson, G. A.Approximation Theory and Numerical Methods, Wiley, Chichester (1980).Google Scholar
  17. 17.
    Yamakawa, E., Fukushima, M. and Ibaraki, T.:An efficient trust region algorithm for minimizing nondifferentiable composite functions, SIAM J Sci Stat Comp 10, (1989) 562–580.Google Scholar
  18. 18.
    Yiu, K. F. C.:Numerical methods for the orthogonal l 1 distance regression problem, M Sc Dissertation, University of Dundee, (1989).Google Scholar

Copyright information

© BIT Foundations 1991

Authors and Affiliations

  • G. A. Watson
    • 1
  • K. F. C. Yiu
    • 1
  1. 1.Department of Mathematics and Computer ScienceUniversity of DundeeDundeeScotland

Personalised recommendations