Communications for Statistical Applications and Methods

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Linear regression under log-concave and Gaussian scale mixture errors: comparative study

  • Kim, Sunyul (Department of Statistics, Sungkyunkwan University) ;
  • Seo, Byungtae (Department of Statistics, Sungkyunkwan University)
  • Received : 2018.06.08
  • Accepted : 2018.09.04
  • Published : 2018.11.30
Download PDF
⟨ Previous Next ⟩

Abstract

Gaussian error distributions are a common choice in traditional regression models for the maximum likelihood (ML) method. However, this distributional assumption is often suspicious especially when the error distribution is skewed or has heavy tails. In both cases, the ML method under normality could break down or lose efficiency. In this paper, we consider the log-concave and Gaussian scale mixture distributions for error distributions. For the log-concave errors, we propose to use a smoothed maximum likelihood estimator for stable and faster computation. Based on this, we perform comparative simulation studies to see the performance of coefficient estimates under normal, Gaussian scale mixture, and log-concave errors. In addition, we also consider real data analysis using Stack loss plant data and Korean labor and income panel data.

Keywords

References

  1. Bellio R and Ventura L (2005). An introduction to robust estimation with R functions. In Proceedings of 1st International Work, 1-57.
  2. Bohning D (1995). A review of reliable maximum likelihood algorithms for semiparametric mixture models, Journal of Statistical Planning and Inference, 47, 5-28. https://doi.org/10.1016/0378-3758(94)00119-G
  3. Brownlee KA (1960). Statistical Theory and Methodology in Science and Engineering, John Wiley & Sons, New York.
  4. Chen Y and Samworth RJ (2013). Smoothed log-concave maximum likelihood estimation with applications, Statistica Sinica, 23, 1373-1398.
  5. Dharmadhikari SW and Joag-Dev K (1988). Unimodality, Convexity, and Applications, Academic Press, Boston.
  6. Dumbgen L, Husler A, and Rufibach K (2007). Active set and EM algorithms for log-concave densities based on complete and censored data, arXiv:0707.4643.
  7. Dumbgen L, Samworth R, and Schuhmacher D (2011). Approximation by log-concave distributions, with applications to regression, The Annals of Statistics, 39, 702-730. https://doi.org/10.1214/10-AOS853
  8. Holland PW and Welsch RE (1977). Robust regression using iteratively reweighted least-squares, Communications in Statistics - Theory and Methods, 6, 813-827. https://doi.org/10.1080/03610927708827533
  9. Karlin S (1968). Total Positivity, Stanford University Press, Stanford.
  10. Lange KL, Little RJA, and Taylor JMG (1989). Robust statistical modeling using the t distribution, Journal of the American Statistical Association, 84, 881-896.
  11. Lange K and Sinsheimer JS (1993). Normal/independent distributions and their applications in robust regression, Journal of Computational and Graphical Statistics, 2, 175-198.
  12. ML and Kalbfleisch JD (1992). An algorithm for computing the nonparametric MLE of a mixing distribution, Journal of the American Statistical Association, 87, 120-126. https://doi.org/10.1080/01621459.1992.10475182
  13. Lindsay BG (1983). The geometry of mixture likelihoods: a general theory, The Annals of Statistics, 11, 86-94. https://doi.org/10.1214/aos/1176346059
  14. Pal JK, Woodroofe M, and Meyer M (2007). Estimating a Polya frequency $function_2$, 54, Lecture Notes-Monograph Series, 239-249.
  15. Rufibach K (2007). Computing maximum likelihood estimators of a log-concave density function, Journal of Statistical Computation and Simulation, 77, 561-574. https://doi.org/10.1080/10629360600569097
  16. Rufibach K and Dumbgen L (2010). Logcondens: estimate a log-concave probability density from iid observations, R package version, 2.
  17. Seo B, Noh J, Lee T, and Yoon YJ (2017). Adaptive robust regression with continuous Gaussian scale mixture errors, Journal of the Korean Statistical Society, 46, 113-125. https://doi.org/10.1016/j.jkss.2016.08.002
  18. Silverman BW (1982). On the estimation of a probability density function by the maximum penalized likelihood method, The Annals of Statistics, 10, 795-810. https://doi.org/10.1214/aos/1176345872
  19. Walther G (2002). Detecting the presence of mixing with multiscale maximum likelihood, Journal of the American Statistical Association, 97, 508-513. https://doi.org/10.1198/016214502760047032
  20. Walther G (2009). Inference and modeling with log-concave distributions, Statistical Science, 24, 319-327. https://doi.org/10.1214/09-STS303
  21. Wang Y (2007). On fast computation of the non-parametric maximum likelihood estimate of a mixing distribution, Journal of the Royal Statistical Society: Series B (Statistical Methodology), 69, 185-198. https://doi.org/10.1111/j.1467-9868.2007.00583.x