Variance intervention

Authors

  • Peter Young,

    Corresponding author
    1. University of Lancaster, UK
    • Centre for Research on Environmental Systems, Institute for Environmental and Biological Sciences, University of Lancaster, LA 1 4YQ, UK
    Search for more papers by this author
    • Peter C. Young is Director of the Centre for Research on Environmental Systems at the University of Lancaster. From 1975 to 1981 he was Professional Fellow at the Australian National University and, between 1970 and 1975, he was University Lecturer in Control and Systems, and Fellow of Clare Hall, University of Cambridge. His research interests include the recursive estimation, forecasting and control of non-stationary dynamic systems, with applications in areas ranging from ecology through engineering to economics.

  • Cho Ng

    Corresponding author
    1. University of Lancaster, UK
    • Centre for Research on Environmental Systems, Institute for Environmental and Biological Sciences, University of Lancaster, LA 1 4YQ, UK
    Search for more papers by this author
    • Cho Nam Ng is Post Doctoral Research Fellow in the Centre for Research on Environmental Systems at the University of Lancaster. He obtained a first degree in Environmental Science in 1983 and a PhD in 1987, both from the University of Lancaster. His research interests include the recursive estimation and forecasting of non-stationary time-series and environmental systems analysis


Abstract

Variance intervention is a simple state-space approach to handling sharp discontinuities of level or slope in the states or parameters of models for non-stationary time-series. It derives from earlier procedures used in the 1960s for the design of self-adaptive, state variable feedback control systems. In the alternative state-space forecasting context considered in the present paper, it is particularly useful when applied to structural time series models. The paper compares the variance intervention procedure with the related ‘subjective intervention’ approach proposed by West and Harrison in a recent issue of the Journal of Forecasting, and demonstrates it efficacy by application to various time-series data, including those used by West and Harrison.

Ancillary