Volume 47 - Article 8 | Pages 199–232

Leveraging deep neural networks to estimate age-specific mortality from life expectancy at birth

By Andrea Nigri, Susanna Levantesi, Josè Manuel Aburto

Print this page  Facebook  Twitter

 

 
Date received:02 Apr 2022
Date published:16 Jul 2022
Word count:6300
Keywords:death rates, deep neural network, forecasting, life expectancy
DOI:10.4054/DemRes.2022.47.8
Additional files:47-8 Additional_Information (pdf file, 11 MB)
 

Abstract

Background: Life expectancy is one of the most informative indicators of population health and development. Its stability, which has been observed over time, has made the prediction and forecasting of life expectancy an appealing area of study. However, predicted or estimated values of life expectancy do not tell us about age-specific mortality.

Objective: Reliable estimates of age-specific mortality are essential in the study of health inequalities, well-being and to calculate other demographic indicators. This task comes with several difficulties, including a lack of reliable data in many populations. Models that relate levels of life expectancy to a full age-specific mortality profile are therefore important but scarce.

Methods: We propose a deep neural networks (DNN) model to derive age-specific mortality from observed or predicted life expectancy by leveraging deep-learning algorithms akin to demography’s indirect estimation techniques.

Results: Out-of-sample validation was used to validate the model, and the predictive performance of the DNN model was compared with two state-of-the-art models. The DNN model provides reliable estimates of age-specific mortality for the United States, Italy, Japan, and Russia using data from the Human Mortality Database.

Contribution: We show how the DNN model could be used to estimate age-specific mortality for countries without age-specific data using neighbouring information or populations with similar mortality dynamics. We take a step forward among demographic methods, offering a multi-population indirect estimation based on a data driven-approach, that can be fitted to many populations simultaneously, using DNN optimisation approaches.

Author's Affiliation

Andrea Nigri - Università Commerciale Luigi Bocconi, Italy [Email]
Susanna Levantesi - Università degli Studi di Roma La Sapienza, Italy [Email]
Josè Manuel Aburto - University of Oxford, United Kingdom [Email]

Other articles by the same author/authors in Demographic Research

» Divergent trends in lifespan variation during mortality crises
Volume 46 - Article 11

» Mexican mortality 1990‒2016: Comparison of unadjusted and adjusted estimates
Volume 44 - Article 30

» Lexis fields
Volume 42 - Article 24

» Geofaceting: Aligning small-multiples for regions in a spatially meaningful way
Volume 41 - Article 17

» The threshold age of the lifetable entropy
Volume 41 - Article 4

Most recent similar articles in Demographic Research

» Point and interval forecasts of mortality rates and life expectancy: A comparison of ten principal component methods
Volume 25 - Article 5    | Keywords: forecasting, life expectancy

» Life expectancy loss among Native Americans during the COVID-19 pandemic
Volume 47 - Article 9    | Keywords: life expectancy

» The contribution of smoking-attributable mortality to differences in mortality and life expectancy among US African-American and white adults, 2000–2019
Volume 46 - Article 31    | Keywords: life expectancy

» To what extent were life expectancy gains in South Africa attributable to declines in HIV/AIDS mortality from 2006 to 2017? A life table analysis of age-specific mortality
Volume 46 - Article 18    | Keywords: life expectancy

» Is the accuracy of individuals' survival beliefs associated with their knowledge of population life expectancy?
Volume 45 - Article 14    | Keywords: life expectancy