Extensions on the Hatzopoulos–Sagianou Multiple-Components Stochastic Mortality Model

In this paper, we present extensions of the Hatzopoulos–Sagianou (2020) (HS) multiple-component stochastic mortality model. Our aim is to thoroughly evaluate and stress test HS model by deploying various link functions, using generalised linear models, diverse distributions in model’s estimation method. work, differentiate approach modelling number deaths Binomial commonly employed literature modelling. Given this, new are derived off-the-shelf namely complementary log–log, logit probit, while also reform introducing a form functions with particular focus on use heavy-tailed distributions. The above-mentioned enhancements conclude methodology for model, prove that it more suitable than those used dynamics. regard, our work offers an extensive experimental testbed scrutinise efficiency, explainability capacity both newly introduced over datasets different characteristics. bring improvement approximately 16% goodness-of-fit, they uncover fine-grained age clusters. addition, compare performance against other well-known under fitting forecast modes. results reflect advantageous features deliver highly informative structure enable attribution identified trend unique cluster. improvements analysts perform in-depth detailed investigation trends specific clusters can contribute attempts academia industry tackle uncertainties risks increasing life expectancy.