Abstract:Factorizing multiparameter densities are proposed for the analytic continuum modeling of human life tables. The formalism is developed based on mortality data of the female, male, and total population of France for the year 2021. The data sets cover the age range from birth up to 110 years. The cumulative hazard function is a multiply broken exponential density, admitting a differential hazard rate capable of describing the observed late-life mortality deceleration and exhibiting exponential asymptotic increase rather than a mortality plateau. The nonlinear least-squares regression is performed with the survival function, which admits double-exponential decay in the high-age limit. The minimization of the multiparametric least-squares functional is facilitated by invoking the product structure of the cumulative hazard, the number of factors and fitting parameters being adapted to the data set. More generally, new techniques are developed to deal with probability densities exhibiting a nonlinear multiparameter dependence. Such densities are increasingly needed to represent extended data sets, as exemplified by recent mortality data across the tree of life. In the case of the mentioned human life tables, the residual deviations of the regressed survival functions and cumulative distributions from the data points are within at most two percent, uniformly over the entire empirical age range. The analytic probability density and age-conditioned survival probabilities are calculated for the total population and the female and male cohorts. The lifetime evolution of the cumulative and differential female/male hazard ratios is studied, from birth up to the asymptotic high-age regime.

Forecasting vital rates by a trimodal extension of the flexible generalized skew normal probability density function

Nonlinear multiparametric modeling of life-table data with adaptive distributions: time evolution of hazard ratios

Robust Non-Parametric Mortality and Fertility Modelling and Forecasting: Gaussian Process Regression Approaches

Multi-population mortality modelling and forecasting with divergence bounds

Probabilistic mortality forecasting with varying age-specific survival improvements

Projecting UK Mortality using Bayesian Generalised Additive Models

Mortality and life expectancy forecasting for a group of populations in developed countries: A multilevel functional data method

Learning and forecasting of age-specific period mortality via B-spline processes with locally-adaptive dynamic coefficients

Robust forecasting of mortality and fertility rates: A functional data approach

Stochastic frailty models for modeling and forecasting mortality

Gompertz Law Revisited: Forecasting Mortality with a Multi-factor Exponential Model

Bayesian forecasting of mortality rates using latent Gaussian models

Deepening Lee-Carter for longevity projections with uncertainty estimation

Bayesian model averaging for mortality forecasting using leave-future-out validation

Mortality modeling at old-age: a mixture model approach

A Mortality Model for Multi-populations: A Semi-Parametric Approach

Multipopulation mortality modelling and forecasting: The multivariate functional principal component with time weightings approaches

Frailty-based mortality models and reserving for longevity risk

Generalized Additive Forecasting Mortality

Gender‐inclusive financial and demographic literacy: Monetizing the gender mortality gap

Mortality Modelling using Generalized Estimating Equations