Boosting Domain-Specific Models with Shrinkage: An Application in Mortality Forecasting

Han LI
Associate Professor
University of Melbourne

Tuesday, 26 March 2024
2:30 PM – 4:00 PM
Venue: Gaia Lecture Theatre 5 (#ABS-02-LT5)

Chairperson: Associate Prof Wenjun Zhu

Abstract

This paper extends the technique of gradient boosting with a focus on using domain-specific models instead of trees. The domain of mortality forecasting is considered as an application. The two novel contributions are to use well-known stochastic mortality models as weak learners in gradient boosting rather than trees, and to include a penalty that shrinks the forecasts of mortality in adjacent age groups and nearby geographical regions closer together. The proposed method demonstrates superior forecasting performance based on US male mortality data from 1969 to 2019. The proposed approach also enables us to interpret and visualize the results. The boosted model with age-based shrinkage yields the most accurate national-level mortality forecast. For state-level forecasts, spatial shrinkage provides further improvement in accuracy in addition to the benefits achieved by age-based shrinkage. This additional improvement can be attributed to data sharing across states with both large and small populations in adjacent regions, as well as states which share common risk factors.

About the Speaker

Dr Han Li is an Associate Professor at the Centre for Actuarial Studies, the University of Melbourne. She is an Associate of the Institute of Actuaries of Australia, and has a broad range of research interests around longevity and mortality risks, ageing and retirement, and the impact of climate change on insurance industry. She has attracted research funds from the Australian Research Council, the Society of Actuaries, and the Casualty Actuarial Society. Han's research has been published in top-tier journals including Insurance: Mathematics and Economics, ASTIN Bulletin, North American Actuarial Journal, Scandinavian Actuarial Journal, Journal of Forecasting, International Journal of Forecasting, and Annals of Actuarial Science.