Kybernetika 48 no. 5, 1045-1063, 2012

Bias correction on censored least squares regression models

Jesus Orbe and Vicente Núñez-Antón

Abstract:

This paper proposes a bias reduction of the coefficients' estimator for linear regression models when observations are randomly censored and the error distribution is unknown. The proposed bias correction is applied to the weighted least squares estimator proposed by Stute [28] [W. Stute: Consistent estimation under random censorship when covariables are present. J. Multivariate Anal. \emph{45} (1993), 89-103.], and it is based on model-based bootstrap resampling techniques that also allow us to work with censored data. Our bias-corrected estimator proposal is evaluated and its behavior assessed in simulation studies concluding that both the bias and the mean square error are reduced with the new proposal.

Keywords:

linear regression, bias, censoring, least squares, Kaplan-Meier estimator

Classification:

62N01, 62F40

paper.pdf

References:

  1. M. G. Akritas: Bootstrapping the Kaplan-Meier estimator. J. Amer. Statist. Assoc. 81 (1986), 1032-1038.   CrossRef
  2. D. G. Altman, B. L. De Stavola, S. B. Love and K. A. Stepniewska: Review of survival analyses published in cancer journals. British J. Cancer. 72 (1985), 511-518.   CrossRef
  3. J. J. Buckley and I. R. James: Linear regression with censored data. Biometrika 66 (1979), 429-436.   CrossRef
  4. S. Chatterjee and D. L. McLeish: Fitting linear regression models to censored data by least squares and maximum likelihood methods. Comm. Statist. Theory Methods 15 (1986), 3227-3243.   CrossRef
  5. Y. Y. Chen, M. Hollander and N. A. Langberg: Small sample results for the Kaplan-Meier estimator. J. Amer. Statist. Assoc. 77 (1982), 141-144.   CrossRef
  6. D. R. Cox: Regression models and life-tables. J. R. Stat. Soc. Ser. B. 34 (1972), 187-220.   CrossRef
  7. D. R. Cox: Partial likelihood. Biometrika 62 (1975), 269-276.   CrossRef
  8. A. C. Davison and D. V. Hinkley: Bootstrap Methods and Their Application. Cambridge University Press, Cambridge 1997.   CrossRef
  9. B. Efron: The two sample problem with censored data. In: Proc. 5th Berkeley Symposium 4 (1967), pp. 831-853.   CrossRef
  10. B. Efron: Censored data and the bootstrap. J. Amer. Statist. Assoc. 76 (1981), 312-319.   CrossRef
  11. B. Efron and R. J. Tibshirani: An Introduction to the Bootstrap. Chapman and Hall, New York 1993.   CrossRef
  12. R. D. Gill: Censoring and Stochastics Integrals. Math. Centre Tracts 124, Math. Centrum, Amsterdam 1980.   CrossRef
  13. G. Heller and J. S. Simonoff: A comparison of estimators for regression with a censored response variable. Biometrika 77 (1990), 515-520.   CrossRef
  14. Z. Jin, D. Lin, L. J. Wei and Z. Ying: Rank-based inference for the accelerated failure time model. Biometrika 90 (2003), 341-353.   CrossRef
  15. E. L. Kaplan and P. Meier: Nonparametric estimation from incomplete observations. J. Amer. Statist. Assoc. 53 (1958), 457-481.   CrossRef
  16. H. Koul, V. Susarla and J. Van-Ryzin: Regression analysis with randomly right-censored data. Ann. Statist. 9 (1981), 1279-1288.   CrossRef
  17. T. L. Lai and Z. Ying: Linear rank statistics in regression analysis with censored or truncated data. J. Multivariate Anal. 40 (1992), 13-45.   CrossRef
  18. S. Leurgans: Linear models, random censoring and synthetic data. Biometrika 74 (1987), 301-309.   CrossRef
  19. D. Mauro: A combinatoric approach to the Kaplan-Meier estimator. Ann. Statist. 13 (1985), 142-149.   CrossRef
  20. R. G. Miller: Least squares regression with censored data. Biometrika 63 (1976), 449-464.   CrossRef
  21. R. G. Miller and J. Halpern: Regression with censored data. Biometrika 69 (1982), 521-531.   CrossRef
  22. J. Orbe, E. Ferreira and V. Núñez-Antón: Censored partial regression. Biostatistics 4 (2003), 109-121.   CrossRef
  23. N. Reid: Estimating the median survival time. Biometrika 68 (1981), 601-608.   CrossRef
  24. N. Reid: A conversation with Sir David Cox. Statist. Sci. 9 (1994), 439-455.   CrossRef
  25. Y. Ritov: Estimation in a linear regression model with censored data. Ann. Statist. 18 (1990), 303-328.   CrossRef
  26. J. Schmee and G. J. Hahn: A simple method for regression analysis with censored data (with discussion). Technometrics 21 (1979), 417-434.   CrossRef
  27. J. Stare, F. Heinzl and F. Harrel: On the use of Buckley and James least squares regression for survival data. In: New Approaches in Applied Statistics (A. Ferligoj and A. Mrvar, eds.), Metodološki zvezki 16, Ljubljana: Eslovenia, 2000, pp. 125-134.   CrossRef
  28. W. Stute: Consistent estimation under random censorship when covariables are present. J. Multivariate Anal. 45 (1993), 89-103.   CrossRef
  29. W. Stute: The bias of Kaplan-Meier integrals. Scand. J. Stat. 21 (1994), 475-484.   CrossRef
  30. W. Stute: Improved estimation under random censorship. Comm. Statist. Theory Methods 23 (1994), 2671-2682.   CrossRef
  31. W. Stute: Distributional convergence under random censorship when covariables are present. Scand. J. Stat. 23 (1996), 461-471.   CrossRef
  32. W. Stute: The jackknife estimate of variance of a Kaplan-Meier integral. Ann. Statist. 24 (1996), 2679-2704.   CrossRef
  33. W. Stute: Nonlinear censored regression. Statist. Sinica 9 (1999), 1089-1102.   CrossRef
  34. W. Stute and J. L. Wang: The jackknife estimate of a Kaplan-Meier integral. Biometrika 81 (1994), 602-606.   CrossRef
  35. A. A. Tsiatis: Estimating regression parameters using linear rank tests for censored data. Ann. Statist. 18 (1990), 354-372.   CrossRef
  36. L. J. Wei: The accelerated failure time model: a useful alternative to the Cox regression model in survival analysis. Stat. Med. 11 (1992), 1871-1879.   CrossRef
  37. J. A. Wellner: A heavy censoring limit theorem for the product limit estimator. Ann. Statist. 13 (1985), 150-162.   CrossRef
  38. M. Zhou: Two-sided bias bound of the Kaplan-Meier estimator. Probab. Theory and Related Fields 79 (1988), 165-173.   CrossRef