Hostname: page-component-cd9895bd7-gvvz8 Total loading time: 0 Render date: 2024-12-27T14:01:48.843Z Has data issue: false hasContentIssue false
  • English
  • Français

Multi-output Gaussian processes for multi-population longevity modelling

Published online by Cambridge University Press:  17 May 2021

Nhan Huynh  and
Mike Ludkovski Open the ORCID record for Mike Ludkovski [Opens in a new window]
Nhan Huynh
Affiliation:
Department of Statistics and Applied Probability, University of California at Santa Barbara, Santa Barbara, CA93106, USA
Mike Ludkovski*
Affiliation:
Department of Statistics and Applied Probability, University of California at Santa Barbara, Santa Barbara, CA93106, USA
*
*Corresponding author. E-mail: [email protected]
Get access Rights & Permissions [Opens in a new window]

Abstract

We investigate joint modelling of longevity trends using the spatial statistical framework of Gaussian process (GP) regression. Our analysis is motivated by the Human Mortality Database (HMD) that provides unified raw mortality tables for nearly 40 countries. Yet few stochastic models exist for handling more than two populations at a time. To bridge this gap, we leverage a spatial covariance framework from machine learning that treats populations as distinct levels of a factor covariate, explicitly capturing the cross-population dependence. The proposed multi-output GP models straightforwardly scale up to a dozen populations and moreover intrinsically generate coherent joint longevity scenarios. In our numerous case studies, we investigate predictive gains from aggregating mortality experience across nations and genders, including by borrowing the most recently available “foreign” data. We show that in our approach, information fusion leads to more precise (and statistically more credible) forecasts. We implement our models in R, as well as a Bayesian version in Stan that provides further uncertainty quantification regarding the estimated mortality covariance structure. All examples utilise public HMD datasets.

Keywords

longevity modelingGaussian processmulti-population mortality
Type
Paper
Information
Annals of Actuarial Science , Volume 15 , Issue 2 , July 2021 , pp. 318 - 345
Copyright
© The Author(s), 2021. Published by Cambridge University Press on behalf of Institute and Faculty of Actuaries

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Alvarez, M.A., Rosasco, L. & Lawrence, N.D. (2011). Kernels for vector-valued functions: a review.CrossRefGoogle Scholar
Antonio, K., Devriendt, S., de Boer, W., de Vries, R., De Waegenaere, A., Kan, H.-K., Kromme, E., Ouburg, W., Schulteis, T., Slagter, E. et al. (2017). Producing the Dutch and Belgian mortality projections: a stochastic multi-population standard. European Actuarial Journal, 7(2), 297–336.CrossRefGoogle Scholar
Armstrong, J.S. & Collopy, F. (1992). Error measures for generalizing about forecasting methods: empirical comparisons. International Journal of Forecasting, 8(1), 69–80.CrossRefGoogle Scholar
Betancourt, M. (2017). Robust Gaussian processes in Stan. Available online at the address https://betanalpha.github.io/assets/case_studies/gp_part3/part3.html.Google Scholar
Boe, C., Winant, C., Riffe, T., Barbieri, M., Wilmoth, J.R., Jasilionis, D., Grigoriev, P., Jdanov, D., Shkolnikov, V.M. & Glei, D. (2015). Data resource profile: The Human Mortality Database (HMD). International Journal of Epidemiology, 44 (5), 1549–1556.Google Scholar
Bonilla, E.V., Chai, K.M. & Williams, C. (2008). Multi-task Gaussian process prediction. In Platt, J.C., Koller, D., Singer, Y. & Roweis, S.T. (Eds.), Advances in Neural Information Processing Systems 20 (pp. 153–160). Curran Associates, Inc.Google Scholar
Boonen, T.J. & Li, H. (2017). Modeling and forecasting mortality with economic growth: a multipopulation approach. Demography, 54(5), 1921–1946.CrossRefGoogle ScholarPubMed
Booth, H. & Tickle, L. (2008). Mortality modelling and forecasting: a review of methods. Annals of Actuarial Science, 3(1–2), 3–43.10.1017/S1748499500000440CrossRefGoogle Scholar
Carpenter, B., Gelman, A., Hoffman, M.D., Lee, D., Goodrich, B., Betancourt, M., Brubaker, M., Guo, J., Li, P. & Riddell, A. (2017). Stan: a probabilistic programming language. Journal of Statistical Software, 76(1).CrossRefGoogle Scholar
Carracedo, P., Debón, A., Iftimi, A. & Montes, F. (2018). Detecting spatio-temporal mortality clusters of European countries by sex and age. International Journal for Equity in Health, 17 (1), 38.CrossRefGoogle ScholarPubMed
Caruana, R. (1997). Multitask learning. Machine Learning, 28(1), 41–75. ISSN 1573-0565. doi: 10.1023/A:1007379606734.CrossRefGoogle Scholar
Chen, H., MacMinn, R. & Sun, T. (2015). Multi-population mortality models: a factor copula approach. Insurance: Mathematics and Economics, 63, 135–146.Google Scholar
Chiles, J.-P. & Delfiner, P. (1999). Geostatistics: Modeling Spatial Uncertainty. Wiley. ISBN 0471083151 9780471083153.CrossRefGoogle Scholar
Christiansen, M.C., Spodarev, E. & Unseld, V. (2015). Differences in European mortality rates: a geometric approach on the age–period plane. ASTIN Bulletin: The Journal of the IAA, 45 (3), 477–502.CrossRefGoogle Scholar
Chu, W. & Ghahramani, Z. (2005). Gaussian processes for ordinal regression. Journal of Machine Learning Research, 6, 1019–1041.Google Scholar
D’Amato, V., Haberman, S., Piscopo, G., Russolillo, M. & Trapani, L. (2016). Multiple mortality modeling in Poisson Lee–Carter framework. Communications in Statistics-Theory and Methods, 45 (6), 1723–1732.CrossRefGoogle Scholar
Debón, A., Martnez-Ruiz, F. & Montes, F. (2010). A geostatistical approach for dynamic life tables: the effect of mortality on remaining lifetime and annuities. Insurance: Mathematics and Economics, 47 (3), 327–336.Google Scholar
Delwarde, A., Denuit, M., Guillén, M. & Vidiella-i Anguera, A. (2006). Application of the Poisson log-bilinear projection model to the G5 mortality experience. Belgian Actuarial Bulletin, 6(1), 54–68.Google Scholar
Deville, Y., Ginsbourger, D. & Roustant, O. (2019). Contributors: Nicolas Durrande. kergp: Gaussian Process Laboratory. Available online at the address https://CRAN.R-project.org/package=kergp. R package version 0.5.0.Google Scholar
Dong, Y., Huang, F., Yu, H. & Haberman, S. (2020). Multi-population mortality forecasting using tensor decomposition. Scandinavian Actuarial Journal, 2020(8), 754–775.CrossRefGoogle Scholar
Duvenaud, D.K. (2014). Automatic Model Construction with Gaussian Processes. PhD thesis, University of Cambridge.Google Scholar
Enchev, V., Kleinow, T. & Cairns, A.J.G. (2017). Multi-population mortality models: fitting, forecasting and comparisons. Scandinavian Actuarial Journal, 2017(4), 319–342.CrossRefGoogle Scholar
Flaxman, S., Gelman, A., Neill, D., Smola, A., Vehtari, A. & Wilson, A.G. (2015). Fast hierarchical Gaussian processes, technical report, Preprint at http://sethrf.com/files/fast-hierarchical-GPs.pdf.Google Scholar
Garrido-Merchán, E.C. & Hernández-Lobato, D. (2018). Dealing with categorical and integer-valued variables in Bayesian optimization with Gaussian processes. CoRR, abs/1805.03463.Google Scholar
Guibert, Q., Lopez, O. & Piette, P. (2019). Forecasting mortality rate improvements with a high-dimensional VAR. Insurance: Mathematics and Economics, 88, 255–272.Google Scholar
HMD. (2018). The Human Mortality Database. University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available online at the address www.mortality.org.Google Scholar
Hoef, J.M.V. & Barry, R.P. (1998). Constructing and fitting models for cokriging and multivariable spatial prediction. Journal of Statistical Planning and Inference, 69 (2), 275–294. ISSN 0378-3758.CrossRefGoogle Scholar
Huynh, N., Ludkovski, M. & Zail, H. (2020). Multi-population longevity models: a spatial random field approach. In Proceedings of the Society of Actuaries 2020 Living to 100 Symposium.Google Scholar
Hyndman, R.J., Booth, H. & Yasmeen, F. (2013). Coherent mortality forecasting: the product-ratio method with functional time series models. Demography, 50(1), 261–283.10.1007/s13524-012-0145-5CrossRefGoogle ScholarPubMed
Kleinow, T. (2015). A common age effect model for the mortality of multiple populations. Insurance: Mathematics and Economics, 63, 147–152.Google Scholar
Kleinow, T. & Cairns, A.J.G. (2013). Mortality and smoking prevalence: an empirical investigation in ten developed countries. British Actuarial Journal, 18(2), 452–466.CrossRefGoogle Scholar
Letham, B. & Bakshy, E. (2019). Bayesian optimization for policy search via online-offline experimentation.Google Scholar
Li, H. & Lu, Y. (2017). Coherent forecasting of mortality rates: A sparse vector-autoregression approach. ASTIN Bulletin: The Journal of the IAA, 47 (2), 563–600.10.1017/asb.2016.37CrossRefGoogle Scholar
Li, J. (2013). A Poisson common factor model for projecting mortality and life expectancy jointly for females and males. Population Studies, 67(1), 111–126.CrossRefGoogle ScholarPubMed
Li, N. & Lee, R. (2005). Coherent mortality forecasts for a group of populations: an extension of the Lee-Carter method. Demography, 42(3), 575–594.CrossRefGoogle ScholarPubMed
Lu, Q., Hanewald, K. & Wang, X. (2019). Bayesian hierarchical multi-population mortality modelling for china’s provinces. SSRN. Available online at the address doi: http://dx.doi.org/10.2139/ssrn.3494491.CrossRefGoogle Scholar
Ludkovski, M., Risk, J. & Zail, H. (2018). Gaussian process models for mortality rates and improvement factors. ASTIN Bulletin: The Journal of the IAA, 48 (3), 1307–1347.CrossRefGoogle Scholar
Makridakis, S. (1993). Accuracy measures: theoretical and practical concerns. International Journal of Forecasting, 9(4), 527–529.CrossRefGoogle Scholar
Qian, P.Z.G., Wu, H. & Jeff Wu, C.F. (2008). Gaussian process models for computer experiments with qualitative and quantitative factors. Technometrics, 50(3), 383–396.CrossRefGoogle Scholar
Qin, C. & Jevtic, P. (2016). Multi-population mortality modelling with levy processes. SSRN. Available online at the address doi: http://dx.doi.org/10.2139/ssrn.2813678.Google Scholar
Raftery, A.E., Li, N., Ševčková, H., Gerland, P. & Heilig, G.K. (2012). Bayesian probabilistic population projections for all countries. Proceedings of the National Academy of Sciences, 109, 13915–13921.CrossRefGoogle Scholar
Rasmussen, C.E. & Williams, C.K.I. (2005). Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning). The MIT Press.CrossRefGoogle Scholar
Roustant, O., Ginsbourger, D. & Deville, Y. (2012). DiceKriging, DiceOptim: two R packages for the analysis of computer experiments by kriging-based metamodeling and optimization. Journal of Statistical Software, 51(1), 1–55.CrossRefGoogle Scholar
Roustant, O., Padonou, E., Deville, Y., Clément, A., Perrin, G., Giorla, J. & Wynn, H.P. (2018). Group kernels for Gaussian process metamodels with categorical inputs. Working Paper or preprint, July 2018. Available online at the address https://hal.archives-ouvertes.fr/hal-01702607.Google Scholar
Shang, H.L. (2016). Mortality and life expectancy forecasting for a group of populations in developed countries: a multilevel functional data method. The Annals of Applied Statistics, 10(3), 1639–1672.CrossRefGoogle Scholar
Skolidis, G. & Sanguinetti, G. (2011). Bayesian multitask classification with Gaussian process priors. IEEE Transactions on Neural Networks, 22 (12), 2011–2021. ISSN 1941-0093.CrossRefGoogle ScholarPubMed
Tsai, C.C.-L. & Wu, A.D. (2020). Incorporating hierarchical credibility theory into modelling of multi-country mortality rates. Insurance: Mathematics and Economics, 91, 37–54. ISSN 0167-6687.Google Scholar
Tsai, C.C.-L. & Zhang, Y. (2019). A multi-dimensional bhlmann credibility approach to modeling multi-population mortality rates. Scandinavian Actuarial Journal, 2019(5), 406–431.CrossRefGoogle Scholar
United Nations. (2011). World Population Prospects: The 2010 Revision, Volume I: Comprehensive Tables. Department of Economic and Social Affairs, Population Division. ST/ESA/SER.A/313.Google Scholar
Wang, C.-W., Yang, S.S. & Huang, H.-C. (2015). Modeling multi-country mortality dependence and its application in pricing survivor index swaps”–a dynamic copula approach. Insurance: Mathematics and Economics, 63, 30–39.Google Scholar
Wang, P., Pantelous, A.A. & Vahid, F. (2020). Multi-population mortality projection: the augmented common factor model with structural breaks. SSRN.Google Scholar
Wiśniowski, A., Smith, P.W.F., Bijak, J., Raymer, J. & Forster, J.J. (2015). Bayesian population forecasting: extending the Lee-Carter method. Demography, 52(3), 1035–1059.CrossRefGoogle ScholarPubMed
Yang, S.S. & Wang, C.-W. (2013). Pricing and securitization of multi-country longevity risk with mortality dependence. Insurance: Mathematics and Economics, 52 (2), 157–169.Google Scholar