TRADE Regression

  • Jan B. Dijkstra
Conference paper

Abstract

TRADE regression is a two-step reweighted least squares method. The name comes from TRim And DElete. The first step involves trimming of the residuals and it makes the method robust against outliers. And the second step (in which only extreme outliers are ignored) restores the statistical efficiency for normal errors.

During the last ten years the topic of high-breakdown methods is a very popular one among the developers of statistical methods. Some examples are Repeated Medians (Siegel, 1982), Least Median of Squares (Rousseeuw, 1984), S-estimators (Rousseeuw and Yohai, 1984), MM-estimators (Yohai, 1987) and τ-estimators (Yohai and Zamar, 1988). These very robust methods are extremely time-consuming and therefore only applicable for models with only a few variables. This is one of the reasons why they are not incorporated in the popular statistical packages.

TRADE regression is a compromise. The method is developed with the following objectives: (1) It must be applicable with the tools that are already present in packages like SAS, SPSS and BMDP, (2) It must be suitable for large datasets and (3) There should be a reasonable balance between robustness and statistical efficiency.

Keywords

Hull Guaran 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Atkinson, A.C. (1985) Plots, transformations and regression Clarendon Press, Oxford.Google Scholar
  2. [2]
    Bickel, P.J. (1973) On some analogues to linear combination of order statistics in the linear model Annals of Statistics (1) 597–616.Google Scholar
  3. [3]
    Cook, R.D. (1977) Detection of influential observations in linear regression Technometrics (19) 15–18.Google Scholar
  4. [4]
    Debets, P, J.B. Dijkstra, G. Eilers, H. van der Knaap, A. Otten, J. Raatgever, D. van der Sluis, G. Stemerdink, A. Verbeek and B. van Zomeren (1989) Regressie diagnostiek in verschillende pakket-ten Kwantitatieve Methoden (32) 95–138.Google Scholar
  5. [5]
    Donoho, D.L. and P.J. Huber (1983) The notion of breakdown point In A Festschrift for Erich Lehmann, BelmontGoogle Scholar
  6. [6]
    Hawkins, D.M, D. Bradu and G.V. Kass (1984) Location of several outliers in multiple regression data using elemental sets Technometrics (26) 197–208.Google Scholar
  7. [7]
    Jaeckel, L.A. (1972) Estimating regression coefficients by minimizing the dispersion of residuals Annals of Mathematical Statistics (5) 1449–1458.Google Scholar
  8. [8]
    Jansen, F.J. (1988) Robuuste regressie analyse Afstudeerverslag Wiskunde, Technische Universiteit EindhovenGoogle Scholar
  9. [9]
    Koenker, R. and G.J. Bassett (1978) Regression quantites Econometrica (46) 33–50.CrossRefGoogle Scholar
  10. [10]
    Rousseeuw, P.J. (1984) Least median of squares regression Journal of the American Statistical Association (79) 871–880CrossRefGoogle Scholar
  11. [11]
    Rousseeuw, P.J. and V. Yohai (1984) Robust regression by means of S-estimators Lecture Notes in Statistics (26) 256–272CrossRefGoogle Scholar
  12. [12]
    Rousseeuw, P.J. and B.C. van Zomeren (1990) Unmasking multivariate outliers and leverage points Journal of the American Statistical Association (85) 633–639.CrossRefGoogle Scholar
  13. [13]
    Rousseeu, P.J. and A.M. Leroy (1987) Robust regression and outlier detection John Wiley and Sons, New York.CrossRefGoogle Scholar
  14. [14]
    Ruppert, D. and R.J. Carroll (1980) Trimmed least squares estimation in the linear model Journal of the American Statistical Association (75) 828–838.CrossRefGoogle Scholar
  15. [15]
    Siegel, A.F. (1982) Robust regression using repeated medians Biometrika (69) 242–244.CrossRefGoogle Scholar
  16. [16]
    Yohai, V.J. (1987) High breakdown-point and high efficiency robust estimates for regression The Annals of Statistics (15) 642–656.CrossRefGoogle Scholar
  17. [17]
    Yohai, VJ. and R.H. Zamar (1988) High breakdown-point estimates of regression by means of minimization of an efficient scale Journal of the American Statistical Association (83) 406–413.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1992

Authors and Affiliations

  • Jan B. Dijkstra
    • 1
  1. 1.Computing CentreEindhoven University of TechnologyEindhovenThe Netherlands

Personalised recommendations