The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

A comparison study on regression with stationary nonparametric autoregressive errors

정상 비모수 자기상관 오차항을 갖는 회귀분석에 대한 비교 연구

  • Yu, Kyusang (Department of Applied Statistics, Konkuk University)
  • 유규상 (건국대학교 응용통계학과)
  • Received : 2015.12.15
  • Accepted : 2015.12.31
  • Published : 2016.02.29
Download PDF
⟨ Previous Next ⟩

Abstract

We compare four methods to estimate a regression coefficient under linear regression models with serially correlated errors. We assume that regression errors are generated with nonlinear autoregressive models. The four methods are: ordinary least square estimator, general least square estimator, parametric regression error correction method, and nonparametric regression error correction method. We also discuss some properties of nonlinear autoregressive models by presenting numerical studies with typical examples. Our numerical study suggests that no method dominates; however, the nonparametric regression error correction method works quite well.

이 논문에서는 비선형 자기회귀 과정을 따르는 오차항을 포함한 회귀모형에서 계수추정법의 비교를 다룬다. 비교를 위해 통상적 최소제곱추정량, 일반화 최소제곱추정량, 모수적 회귀오차 수정법, 비모수적 회귀오차 추정법을 비교하였다. 본 논문에서는 또한 비선형 자기회귀모형의 성질을 전형적인 몇가지 비선형자기회귀 모형을 예를 들어 설명한다. 비교연구의 결과 네 가지 추정량 중에 모든 상황에서 최선인 추정량은 존재하지 않았으나 비모수 회귀오차 수정 방법이 일반적으로 우수한 성능을 보임을 알 수 있다.

Keywords

References

  1. Bhattacharya, R. and Lee, C. (1995). Ergodicity of nonlinear first order autoregressive models, Journal of Theoretical Probability, 8, 210-219.
  2. Biscay, R. J., Lavielle, M., and Ludena, C. (2005). Estimation of nonparametric autoregressive time series models under dynamical constraints, Journal of Time Series Analysis, 26, 371-397. https://doi.org/10.1111/j.1467-9892.2004.00407.x
  3. Fan, J. and Gijbels, I. (1996). Local Polynomial Modelling and Its Applications, Chapman & Hall, London.
  4. Haggan, V. and Ozaki, T. (1981). Modelling nonlinear random vibrations using an amplitude dependent autoregressive time series model, Biometrika, 68, 189-196. https://doi.org/10.1093/biomet/68.1.189
  5. Sheather, S. (2009). A Modern Approach to Regression with R, Springer, New York.
  6. Su, L. and Ullah, A. (2006). More efficient estimation in nonparametric regression with nonparametric autocorrelated errors, Econometric Theory, 22, 98-126.
  7. Tong, H. (1990). Nonlinear Time Series: A Dynamical Approach, Oxford University Press, Oxford.
  8. Tong, H. and Lim, K. (1980). Threshold autoregression, limit cycles and cyclical data (with Discussion), Journal of Royal Statistical Society B, 42, 245-292
  9. Truong, Y. K. and Stone, C. (1992). Nonparametric function estimation involving time series, The Annals of Statistics, 20, 77-97. https://doi.org/10.1214/aos/1176348513