A Data-Cleaning Augmented Kalman Filter for Robust Estimation of State Space Models

30 Pages Posted: 31 Mar 2016

See all articles by Martyna Marczak

Martyna Marczak

University of Hohenheim

Tommaso Proietti

University of Rome II - Department of Economics and Finance

Stefano Grassi

University of Kent - Canterbury Campus

Date Written: March 29, 2016

Abstract

This article presents a robust augmented Kalman filter that extends the data – cleaning filter (Masreliez and Martin, 1977) to the general state space model featuring nonstationary 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. We investigate the performance of the robust AKF 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.

Keywords: robust filtering, augmented Kalman filter, structural time series model, additive outlier, innovation outlier

JEL Classification: C32, C53, C63

Suggested Citation

Marczak, Martyna and Proietti, Tommaso and Grassi, Stefano, A Data-Cleaning Augmented Kalman Filter for Robust Estimation of State Space Models (March 29, 2016). CEIS Working Paper No. 374. Available at SSRN: https://ssrn.com/abstract=2756074 or http://dx.doi.org/10.2139/ssrn.2756074

Martyna Marczak

University of Hohenheim ( email )

Schloss, Museumsfluegel
Stuttgart, 70593
Germany

HOME PAGE: http://labour.uni-hohenheim.de/

Tommaso Proietti (Contact Author)

University of Rome II - Department of Economics and Finance ( email )

Via Columbia, 2
Rome, 00133
Italy

Stefano Grassi

University of Kent - Canterbury Campus ( email )

Keynes College
Canterbury, Kent CT2 7NP
United Kingdom

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
51
Abstract Views
456
rank
398,672
PlumX Metrics