A data-cleaning augmented Kalman filter for robust estimation of state space models

Martyna Marczak*, Tommaso Proietti, Stefano Grassi

*Corresponding author for this work

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaperJournal articleResearchpeer-review

189 Downloads (Pure)

Abstract

A robust augmented Kalman filter (AKF) is presented for the general state space model featuring non-stationary and regression effects. The robust filter shrinks the observations towards their one-step-ahead prediction based on the past, by bounding the effect of the information carried by a new observation according to an influence function. When maximum likelihood estimation is carried out on the replacement data, an M-type estimator is obtained. The performance of the robust AKF is investigated in two applications using as a modeling framework the basic structural time series model—a popular unobserved components model in the analysis of seasonal time series. First, a Monte Carlo experiment is conducted in order to evaluate the comparative accuracy of the proposed method for estimating the variance parameters. Second, the method is applied in a forecasting context to a large set of European trade statistics series.

Original languageEnglish
JournalEconometrics and Statistics
Volume5
Issue1
Pages (from-to)107-123
Number of pages17
ISSN2468-0389
DOIs
Publication statusPublished - 1 Jan 2018

Keywords

  • Additive outlier
  • Augmented Kalman filter
  • DISTRIBUTIONS
  • Innovation outlier
  • OUTLIER DETECTION
  • Robust filtering
  • SCORING RULES
  • Structural time series model

Fingerprint

Dive into the research topics of 'A data-cleaning augmented Kalman filter for robust estimation of state space models'. Together they form a unique fingerprint.

Cite this