Communications for Statistical Applications and Methods

The Korean Statistical Society (한국통계학회)
권호 표지

Constrained $L_1$-Estimation in Linear Regression

  • Kim, Bu-Yong (Department of Statistics, Sookmyung Women's University)
  • Published : 1998.12.01
Download PDF
⟨ Previous Next ⟩

Abstract

An algorithm is proposed for the $L_1$-estimation with linear equality and inequality constraints in linear regression model. The algorithm employs a linear scaling transformation to obtain the optimal solution of linear programming type problem. And a special scheme is used to maintain the feasibility of the updated solution at each iteration. The convergence of the proposed algorithm is proved. In addition, the updating and orthogonal decomposition techniques are employed to improve the computational efficiency and numerical stability.

Keywords

References

  1. SIAM Journal on Numerical Analysis v.14 An Algorithm for a Restricted Discrete Approximation Problem in the L₁-Norm Armstrong, R. D.;Hultz, J. W.
  2. Communications in Statistics-Simula. Computa. v.6 Algorithms for Restricted Least Absolute Value Estimation Barrodale, I.;Roberts, F. D. K.
  3. SIAM Journal on Numerical Analysis v.15 An Efficient Algorithm for Discrete L₁Linear Approximation with linear Constraints Barrodale, I.;Roberts, F. D. K.
  4. ACM Transactions on Mathematical Software v.6 Linearly Constrained Discrete L₁Problems Bartles, R. H.;Conn, A. R.
  5. Journal of the American Statistical Association v.73 Asymptotic Theory of Least Absolute Error Regression Bassett, G.;Koenker, R.
  6. Alternative Methods of Regression Birkes, D.;Dodge, Y.
  7. Econometrica v.39 Regression with Non-Gaussian Stable Distrurbances ; Some Sampling Results Blattberg, R.;Sargent, T.
  8. Least Absolute Deviations : Theory, Applications, and Algorithms Bloomfield, P.;Steiger, W. L.
  9. Communications in statistics - Theory Meth. v.22 On a necessary Condition for the Consistency of the L₁-estimates in Linear Regression Models Chen, X. R.;Wu, Y.
  10. TIMS/Studies in the Management Sciences v.19 Least Absolute Value Estimation in Linear Regression ; A Review Dielman, T.;Pfaffenberger, R.
  11. Journal of the Korean Statistical Society v.16 A Robust Estimation Procedure for the Linear Regression Model Kim, B. Y.
  12. The Korean Communications in Statistics v.2 On the Robustness of L₁-Estimator in Linear Regression Model Kim, B. Y.
  13. Absolute Deviations Curve Fitting ; An Alternative to Least Squares, Contributions to Survey Sampling and Applied Statistics Pfaffenberger, R. C.;Dinkel, J. J.;H. A. David(ed.)
  14. Rosenberg, B.;Carlson, D.
  15. Communications in Statistics - Simulation and Computation v.B6 A Simple Approximation of the Sampling Distribution of Least Absolute Residuals Regression Estimates Rosenberg, B.;Carlson, D.