The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지

Detection of local structural chages in time series

시계열에서 국소구조변화의 탐지에 관한 연구

  • Jae June Lee (Department of Statistics, Inha University, 253 Yonghyun-dong, Nam-ku, Incheon 402-751, Korea)
  • Published : 1994.09.01
Download PDF
⟨ Previous Next ⟩

Abstract

In time series data, atypical observations are not rare. Several approaches have been proposed to detect a single outlier, but the effectiveness of those procedures is in doubt when patchy outliers are present. In this paper, the atypicality in patchy outliers is interpreted as a local structural change, and a model is introduced to entertain its effect on the series. Based on this model, a statistic and a procedure are proposed for identifying those local structural changes. The performance of the proposed procedure is evaluated through simulation study and the analysis of real data sets.

시계열 자료에서 우리는 이상 관측자료들을 흔히 발견하게 된다. 한 점의 이상 관측자료를 탐지하는 방법은 여러가지가 소개되었지만 연속적인 시점에서 이상자료가 존재하는 경우에 기존의 기법은 적절하지 못한 면이 있다. 이 논문에서는 그러한 자료들을 국소구조변화의 결과로 해석하고 그 변화의 크기를 모형화하는 방법을 제시하였다. 이 모형을 이용하여 그러한 국소구조변화를 탐지할 수 있는 통계량과 탐지과정을 제안하였다. 모의실험과 실제 자료의 분석을 수행하여 제안된 기법의 유용성을 평가하였다.

Keywords

References

  1. Time Series Analysis: forecasting and control(2nd ed.) Box,G.E.P.;Jenkins,G.
  2. Journal of the American Statistical Association v.70 Intervention analysis with applications to economic and environmemtal problems Box,G.E.P.;Tiao,G.C.
  3. Journal of the Royal Statistical Society, Series B v.51 Leave-k-out diagonostics for time series (with discussion) Bruce,A.G.;Martin,R.D.
  4. Technometrics v.30 Estimation of time series parameters in the presence of outliers Chang,I.;Tiao,G.C.;Chen,C.
  5. Unpublished Ph. D. dissertation, University of Wisconsin, Department of Statistics Outliers in Time Series Chang,I.
  6. Sankhya, A v.43 Asmptotic theory for time series containing missing and amplitude modulated observations Dunsmuir,W.;Robinson,P.M.
  7. Journal of the Royal Statistical Society, Series B v.32 Outliers in time series Fox,A.J.
  8. Introduction to Statistical Time Series Fuller,F.R.
  9. Journal of the American Statistical Association v.79 Estimating missing observations in economic time series Harvey,A.C.;Pierse,R.G.
  10. Annals of Statistics v.12 Infinitesimal robustness for autoregressive process Kunsch,H.
  11. Journal of Time Series Analysis v.11 Outlier diagnostics in time series analysis Ledolter,J.
  12. Unpublished Ph. D. dissertation, University of Wisconsin, Department of Statistics Outlier diagnostics in time series analysis Ledolter,J.
  13. Annals of Statistics v.14 Influence functionals for time series (with discussion) Martin,R.D.;Yohai,V.J.
  14. Robust methods for time series, in Applied Time Series Analysis Ⅱ Martin,R.D.;Findley,D.F.(ed.)
  15. Journal of Business nad Economic Statistics v.8 Influential observations in time series Pena,D.
  16. Journal of Time Series Analysis v.10 Estimation and Interpolation of missing values of a stationary time series Pourahmadi,M.
  17. Journal of the American Statistical Association v.81 Time series model specification in the presence of outliers Tsay,R.S.
  18. Journal of Forecasting v.7 Outliers, level shifts, and variance changes in time series