Two New Robust Methods for Time Series
We illustrate our method with a simple time series structural model approach. However, the general approach applies to more sophisticated structural and ARIMA models. This modelling approach is a generalization of the Gaussian mixture modeling of Harrison and Stevens (1976), Smith and West (1983), and West and Harrison (1989).
KeywordsGaussian Mixture Modeling ARIMA Model Level Shift Breakdown Point Tree Pruning
Unable to display preview. Download preview PDF.
- Bruce and Martin (1992). “Tree Based Robust Bayesian Estimation of Time Series Structural Models.” Univ. of Washington, Department of Statistics.Google Scholar
- Dempster, A.P., Laird, N.M., and Rubin, D.B. (1977). “Maximum likelihood from incomplete data via the EM algorithm.” Jour. Roy. Statist Soc., Series B., 39(1):1–38.Google Scholar
- Harvey, A. (1989). Forecastin, structural time series models and the Kaiman filter. Cambridge Univ. Press, Cambridge.Google Scholar
- Harrison, P.J. and Stevens, C.F. (1976). “Bayesian Forecasting” Journal of the Roy. Statist. Soc., Series B, 38:205–247.Google Scholar
- Martin, R.D. and Yohai, V.J. (1992). “Highly robust estimation of autoregressions”. Technical Report, Dept. of Statistics, Univ. of Washington, Seattle, WA.Google Scholar