Ensemble interval forecasts of mortality

From MaRDI portal





This paper aims to study the application of ensemble interval forecasts in mortality modelling. The key idea of ensemble interval forecasts is to combine the prediction intervals obtained from different mortality models with a view to improving the interval coverage. A variety of combination methods are considered. These include the simple and weighted averaging, median, mode, envelope, interior and exterior trimming, probability averaging of endpoints, as well as simple averaging of midpoints. The authors consider three classes of mortality models including the Lee-Carter and Cairns-Blake-Dowd families of models, as well as autoregressive models. Empirical studies are performed based on the mortality data of six populations. It is found that in general, forecasts obtained from the proposed ensemble approach outperform those obtained from a single mortality model. Section 2 of the paper discusses different ensemble interval forecasts. Numerical illustrations based on the mortality data from various populations are provided in Section 3. The final section gives some concluding remarks.











This page was built for publication: Ensemble interval forecasts of mortality

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6931694)