Advertisement

EM-Based Inference for Cyclostationary Time Series with Missing Observations

  • Christiana Drake
  • Oskar KnapikEmail author
  • Jacek Leśkow
Chapter
Part of the Lecture Notes in Mechanical Engineering book series (LNME)

Abstract

Periodically correlated (or cyclostationary) time series are becoming more and more popular in many areas (see Gardner et al. 2006). However, in many practical situations data that can be modeled with such time series is incomplete. Some preliminary results on that problem have been presented in a previous work by Drake et al. (2013) by the authors. In this chapter we propose a new ECM-type algorithm based on conditional likelihood and profile likelihood to extend estimation to the case when observations are missing completely at random.

Keywords

Mean Square Error Profile Likelihood Full Likelihood Conditional Maximum Likelihood Lower Mean Square Error 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Notes

Acknowledgments

The chapter was written while Oskar Knapik was visiting the Department of Statistics at University of California at Davis, USA. Its kind hospitality is greatly appreciated. Oskar Knapik gratefully acknowledges the Kosciuszko Foundation for financial support for this research.

References

  1. Brockwell PJ, Davis RA (2002) Introduction to time series and forecasting. Springer, New YorkGoogle Scholar
  2. Dempster AP, Laird NM, Rubin DB (1977) Maximum likelihood from incomplete data via the EM algorithm (with discussion). J Roy Statist Soc B 39:1–38zbMATHMathSciNetGoogle Scholar
  3. Drake C, Knapik O, Leśkow L (2013) Missing data analysis in cyclostationary model, Technical Transactions (accepted)Google Scholar
  4. Gardner WA, Napolitano A, Paura L (2006) Cyclostationarity: half a century of research. Sig Process 86(4):639–697CrossRefzbMATHGoogle Scholar
  5. Ghogho M, Garel B (1999) Maximum likelihood estimation of amplitude-modulated time series. Sig Process 75:99–116CrossRefzbMATHGoogle Scholar
  6. Hamilton JD (1994) Time series analysis. Princeton University Press, New JerseyGoogle Scholar
  7. Little RJA, Rubin DB (2002) Statistical analysis with missing data, 2nd edn. Wiley, New YorkGoogle Scholar
  8. McLachlan GJ, Krishnan T (2008) The EM algorithm and extensions. 2nd edn. Wiley, New YorkGoogle Scholar
  9. Meng X-L, Rubin DB (1993) Maximum likelihood estimation via the ECM algorithm: a general framework. Biometrika 80(2):267–278CrossRefzbMATHMathSciNetGoogle Scholar
  10. Niu X (1996) Seasonal time series models with periodic variances. Technical report. http://stat.fsu.edu/techreports/M911.pdf
  11. Noomene R (2007) Procedures of Parameters’ estimation of AR(1) models into lineal state-space models. Proceedings of the world congress on engineering 2007 Vol II WCE 2007, London, 2–4 July 2007. http://www.iaeng.org/publication/WCE2007/WCE2007_pp995-999.pdf
  12. Pawitan Y (2001) In all likelihood: statistical modelling and inference using likelihood. Oxford University Press, OxfordGoogle Scholar
  13. Schaffer JL (1997) Analysis of incomplete multivariate data. Chapman & Hall/CRCGoogle Scholar
  14. Stefanakos ChN, Athanassoulis GA (2001) A unified methodology for the analysis, completion and simulation of nonstationary time series with missing values, with application to wave data. Appl Ocean Res 23(4):207–220CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Department of Statistics at University of California at DavisDavisUSA
  2. 2.CREATESDepartment of Economics and Business at Aarhus UniversityAarhusDenmark
  3. 3.Institute of Mathematics at Cracow University of TechnologyKracowPoland

Personalised recommendations