A NEURAL-NETWORK ANALYZER FOR MORTALITY FORECAST

This article proposes a neural-network approach to predict and simulate human mortality rates. This semi-parametric model is capable to detect and duplicate non-linearities observed in the evolution of log-forces of mortality. The method proceeds in two steps. During the first stage, a neural-network-based generalization of the principal component analysis summarizes the information carried by the surface of log-mortality rates in a small number of latent factors. In the second step, these latent factors are forecast with an econometric model. The term structure of log-forces of mortality is next reconstructed by an inverse transformation. The neural analyzer is adjusted to French, UK and US mortality rates, over the period 1946–2000 and validated with data from 2001 to 2014. Numerical experiments reveal that the neural approach has an excellent predictive power, compared to the Lee–Carter model with and without cohort effects.

[1]  G. Lewicki,et al.  Approximation by Superpositions of a Sigmoidal Function , 2003 .

[2]  Donatien Hainaut,et al.  Multidimensional Lee–Carter model with switching mortality processes , 2012 .

[3]  Dong Dong,et al.  Nonlinear principal component analysis-based on principal curves and neural networks , 1994, Proceedings of 1994 American Control Conference - ACC '94.

[4]  Gareth Peters,et al.  Cohort effects in mortality modelling: a Bayesian state-space approach , 2017, Annals of Actuarial Science.

[5]  Roland Baddeley,et al.  Nonlinear principal components analysis of neuronal spike train data , 1997, Biological Cybernetics.

[6]  Ronald Lee.  The Lee-Carter Method for Forecasting Mortality, with Various Extensions and Applications , 2000 .

[7]  I. Currie.  On fitting generalized linear and non-linear models of mortality , 2016 .

[8]  Vladimir K. Kaishev,et al.  Dependent competing risks: Cause elimination and its impact on survival , 2013 .

[9]  M. Kramer.  Nonlinear principal component analysis using autoassociative neural networks , 1991 .

[10]  Gareth Peters,et al.  Stochastic Period and Cohort Effect State-Space Mortality Models Incorporating Demographic Factors via Probabilistic Robust Principle Components , 2017 .

[11]  Christos H. Skiadas,et al.  Forecasting Mortality Rate Using a Neural Network with Fuzzy Inference System , 2008 .

[12]  Paolo Emilio Puddu,et al.  Artificial neural networks versus proportional hazards Cox models to predict 45-year all-cause mortality in the Italian Rural Areas of the Seven Countries Study , 2012, BMC Medical Research Methodology.

[13]  P. McNelis.  Neural networks in finance , 2004 .

[14]  Colin O'Hare,et al.  Explaining young mortality , 2012 .

[15]  Katrien Antonio,et al.  The impact of multiple structural changes on mortality predictions , 2014 .

[16]  Michel Denuit,et al.  Modelling Longevity Dynamics for Pensions and Annuity Business , 2009 .

[17]  Paolo Emilio Puddu,et al.  Artificial neural network versus multiple logistic function to predict 25-year coronary heart disease mortality in the Seven Countries Study , 2009, European journal of cardiovascular prevention and rehabilitation : official journal of the European Society of Cardiology, Working Groups on Epidemiology & Prevention and Cardiac Rehabilitation and Exercise Physiology.

[18]  Gareth W. Peters,et al.  A unified approach to mortality modelling using state-space framework: characterisation, identification, estimation and forecasting , 2016, Annals of Actuarial Science.

[19]  Ronald Lee,et al.  Modeling and forecasting U. S. mortality , 1992 .

[20]  Modeling longevity risks using a principal component approach: A comparison with existing stochastic mortality models , 2010 .

[21]  Steven Haberman,et al.  Lee–Carter mortality forecasting with age-specific enhancement , 2003 .

[22]  Katrien Antonio,et al.  Bayesian Poisson log-bilinear models for mortality projections with multiple populations , 2015 .

[23]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[24]  Steven Haberman,et al.  A cohort-based extension to the Lee-Carter model for mortality reduction factors , 2006 .

[25]  Steven Haberman,et al.  Projecting Mortality Trends , 2004 .

[26]  A. Menotti,et al.  Lifetime competing risks between coronary heart disease mortality and other causes of death during 50years of follow-up. , 2017, International journal of cardiology.

[27]  Adam H. Monahan,et al.  Nonlinear Principal Component Analysis by Neural Networks: Theory and Application to the Lorenz System , 2000 .

[28]  Ermanno Pitacco,et al.  Survival models in a dynamic context: a survey , 2004 .

[29]  A State-Space Estimation of the Lee-Carter Mortality Model and Implications for Annuity Pricing , 2015, 1508.00322.

[30]  P. McNelis.  Neural networks in finance : gaining predictive edge in the market , 2005 .

[31]  Kevin Dowd,et al.  Modelling and management of mortality risk: a review , 2008 .

[32]  D. Cox.  Regression Models and Life-Tables , 1972 .

[33]  David R. Cox,et al.  Regression models and life tables (with discussion , 1972 .

[34]  Michel Denuit,et al.  A Poisson log-bilinear regression approach to the construction of projected lifetables , 2002 .

[35]  E. C. Malthouse,et al.  Limitations of nonlinear PCA as performed with generic neural networks , 1998, IEEE Trans. Neural Networks.