Forecasting Mortality Rate by Singular Spectrum Analysis

Abstract

Singular spectrum analysis (SSA) is a relatively new and powerful non-parametric time series analysis technique that has demonstrated its capability in forecasting different time series in various disciplines. In this paper, we study the feasibility of using the SSA to perform mortality forecasts. Comparisons are made with the Hyndman–Ullah model, which is a new powerful tool in the field of mortality forecasting, and will be considered as a benchmark to evaluate the performance of the SSA for mortality forecasting. We use both SSA and Hyndman–Ullah models to obtain 10 forecasts for the period 2000–2009 in nine European countries including Belgium, Denmark, Finland, France, Italy, The Netherlands, Norway, Sweden and Switzerland. Computational results show a superior accuracy of the SSA forecasting algorithms, when compared with the Hyndman–Ullah approach.