Advertisement

Statistics and Computing

, Volume 27, Issue 6, pp 1617–1637 | Cite as

Segmental dynamic factor analysis for time series of curves

  • Allou Samé
  • Gérard Govaert
Article

Abstract

A new approach is introduced in this article for describing and visualizing time series of curves, where each curve has the particularity of being subject to changes in regime. For this purpose, the curves are represented by a regression model including a latent segmentation, and their temporal evolution is modeled through a Gaussian random walk over low-dimensional factors of the regression coefficients. The resulting model is nothing else than a particular state-space model involving discrete and continuous latent variables, whose parameters are estimated across a sequence of curves through a dedicated variational Expectation-Maximization algorithm. The experimental study conducted on simulated data and real time series of curves has shown encouraging results in terms of visualization of their temporal evolution and forecasting.

Keywords

Functional time series Visualization and forecasting Mixture of regressions Dynamic factor analysis Variational EM Condition monitoring 

Notes

Acknowledgments

The authors wish to thank M. Marc Antoni of SNCF for the data he provided and for the support provided.

References

  1. Baek, J., McLachlan, G.J., Flack, L.: Mixtures of factor analyzers with common factor loadings: applications to the clustering and visualization of high-dimensional data. IEEE Trans. Pattern Anal. Mach. Intell. 32, 1298–1309 (2010)CrossRefGoogle Scholar
  2. Bai, J., Wang, P.: Identification and bayesian estimation of dynamic factor models. J. Bus. Econ. Stat. 33(2), 221–240 (2015)MathSciNetCrossRefGoogle Scholar
  3. Beal, M.J.: Variational algorithms for approximate bayesian inference. PhD thesis, Gatsby Computational Neuroscience Unit, University College London (2003)Google Scholar
  4. Besse, P.C., Cardot, H., Stephenson, D.B.: Autoregressive forecasting of some functional climatic variations. Scand. J. Stat. 27, 673–687 (2000)CrossRefMATHGoogle Scholar
  5. Blei, D.M., Lafferty, J.D.: Dynamic topic models. In: Proceedings of the 23rd International Conference on Machine Learning, ICML ’06, pp. 113–120. ACM, New York (2006)Google Scholar
  6. Chamroukhi, F., Samé, A., Govaert, G., Aknin, P.: A regression model with a hidden logistic process for feature extraction from time serie. In: IEEE International Joint Conference on Neural Networks (IJCNN 2009), Atlanta (2009a)Google Scholar
  7. Chamroukhi, F., Samé, A., Govaert, G., Aknin, P.: Time series modeling by a regression approach based on a latent process. Neural Netw. 22, 593–602 (2009b)CrossRefMATHGoogle Scholar
  8. Dempster, A.P., Laird, N.M., Rubin, D.B.: Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B 39, 1–38 (1977)MathSciNetMATHGoogle Scholar
  9. Diebold, F.X., Li, C.: Forecasting the term structure of government bond yields. J. Econom. 130(2), 337–364 (2006)MathSciNetCrossRefMATHGoogle Scholar
  10. Fearnhead, P., Liu, Z.: On-line inference for multiple changepoint problems. J. R. Stat. Soc. B 69, 589–605 (2007)MathSciNetCrossRefGoogle Scholar
  11. Ferraty, F., Vieu, P.: Nonparametric Functional Data Analysis. Springer, New York (2006)MATHGoogle Scholar
  12. Gareth, J., Hastie, T., Sugar, C.: Principal component models for sparse functional data. Biometrika 87, 587–602 (2000)MathSciNetCrossRefMATHGoogle Scholar
  13. Garnett, R., Osborne, M.A., Reece, S., Rogers, A., Roberts, S.J.: Sequential bayesian prediction in the presence of changepoints and faults. Comput. J. 53(9), 1430–1446 (2010)CrossRefGoogle Scholar
  14. Geweke, J.F.: The dynamic factor analysis of economic time series models. In: Aigner, D.J., Goldberger, A.S. (eds.) Latent Variables in Socio-economic Models, pp. 365–383. North-Holland, Amsterdam (1977)Google Scholar
  15. Green, P.: Iteratively reweighted least squares for maximum likelihood estimation, and some robust and resistant alternatives. J. R. Stat. Soc. Ser. B 46(2), 149–192 (1984)MathSciNetMATHGoogle Scholar
  16. Harvey, A.C.: Forecasting Structural Time Series Models and the Kalman Filter. Cambridge University Press, Cambridge (1989)Google Scholar
  17. Hays, S., Shen, H., Huang, J.Z.: Functional dynamic factor models with application to yield curve forecasting. Ann. Appl. Stat. 6(3), 870–894 (2012)MathSciNetCrossRefMATHGoogle Scholar
  18. Hyndman, R.J., Ullah, S.: Robust forecasting of mortality and fertility rates: a functional data approach. Comput. Statist. Data Anal. 51(10), 4942–4956 (2007)MathSciNetCrossRefMATHGoogle Scholar
  19. Jiang, W., Tanner, M.A.: On the identifiability of mixtures-of-experts. Neural Netw. 12(9), 1253–1258 (1999)CrossRefGoogle Scholar
  20. Jolliffe, I.T.: Principal Component Analysis, 2nd edn. Springer, New York (2002)MATHGoogle Scholar
  21. Jordan, M.I., Ghahramani, Z., Jaakkola, T.S., Saul, L.K.: An introduction to variational methods for graphical models. Mach. Learn. 37(2), 183–233 (1999)CrossRefMATHGoogle Scholar
  22. Kalman, R.E.: A new approach to linear filtering and prediction problems. J. Basic Eng. 82(1), 35–45 (1960)CrossRefGoogle Scholar
  23. Lawley, D.N.: The estimation of factor loadings by the method of maximum likelihood. Proc. R. Soc. Edinb. 60(01), 64–82 (1940)MathSciNetCrossRefMATHGoogle Scholar
  24. McLachlan, G.J., Krishnan, T.: The EM Algorithm and Extensions. Wiley, New York (2008)CrossRefMATHGoogle Scholar
  25. McLachlan, G.J., Peel, D., Bean, R.W.: Modelling high-dimensional data by mixtures of factor analyzers. Computat. Stat. Data Anal. 41(3–4), 379–388 (2003)MathSciNetCrossRefMATHGoogle Scholar
  26. Molaneaar, P.C.: A dynamic factor model for the analysis of multivariate time series. Psychometrika 50, 181–202 (1985)Google Scholar
  27. Neal, R.M.: Probabilistic inference using Markov chain Monte Carlo methods. Technical Report CRG-TR-93-1, Dept. of Computer Science, University of Toronto (1993)Google Scholar
  28. Neal, R., Hinton, G.: A view of the EM algorithm that justifies incremental, sparse, and other variants. In: I, J.M. (ed.) Learning in Graphical Models. Kluwer, Dordrecht (1998)Google Scholar
  29. Pearson, K.: On lines and planes of closest fit to systems of points in space. Philos. Mag. 6(2), 559–572 (1901)CrossRefMATHGoogle Scholar
  30. Ramsay, J.O., Silverman, B.W.: Functional data analysis, 2nd edn. Springer, New York (1997)CrossRefMATHGoogle Scholar
  31. Saatçi, Y., Turner, R., Rasmussen, C.E.: Gaussian process change point models. In: Proceedings of the 27th International Conference on Machine Learning, pp. 927–934, Haifa (2010)Google Scholar
  32. Sarkka, S., Nummenmaa, A.: Recursive noise adaptive kalman filtering by variational bayesian approximations. IEEE Trans. Autom. Control 54(3), 596–600 (2009)MathSciNetCrossRefMATHGoogle Scholar
  33. Shumway, R.H., Stofer, D.S.: An approach to time series smoothing and forecasting using the EM algorithm. J. Time Ser. Anal. 3, 253–264 (1982)CrossRefMATHGoogle Scholar
  34. Shumway, R.H., Stoffer, D.S.: Time Series Analysis and Its Applications. Springer, New York (2011)CrossRefMATHGoogle Scholar
  35. Thomas, L.H.: Elliptic problems in linear differential equations over a network. Technical report, Columbia University (1949)Google Scholar
  36. Varah, J.M.: On the solution of block-tridiagonal systems arising from certain finite-difference equations. Math. Comput. 26(120), 859–868 (1972)MathSciNetCrossRefMATHGoogle Scholar
  37. Zuur, A.F., Fryer, R.J., Jolliffe, I.T., Dekker, R., Beukema, J.J.: Estimating common trends in multivariate time series using dynamic factor analysis. Environmetrics 14, 665–685 (2003)CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  1. 1.Université Paris-Est, IFSTTAR, COSYS, GRETTIAMarne-la-ValléeFrance
  2. 2.Laboratoire Heudiasyc, UMR CNRS 7253Université de Technologie de CompiègneCompiègneFrance

Personalised recommendations