Hierarchical Population Stochastic Mortality Model of Multi-Dimension and Its Application
Stochastic mortality forecasting is the basis for pension managers and insurance compa-nies to measure longevity risk.Modeling multi-population stochastic is the forefront and hotspot of current research.This paper extends the traditional mortality model under a single dimension to a multi-dimensional model combining gender population and regional population,so as to improve the sys-tematicness of the multi-population stochastic mortality model.At the same time,prior parameters are introduced to construct a multi-dimensional hierarchical mortality model,and data cloning technology and Markov Chain Monte Carlo simulation are used to improve the parameter estimation method,so as to make up for the problems of strong subjectivity of prior assumptions and unstable fitting results in traditional methods.Through comparative research,it is found that the multi-dimensional hierarchical stochastic mortality model improves the scientificity of mortality modeling and the rationality of predic-tion results.It not only obtains better in sample goodness of fit and short-term prediction effect,but also the long-term prediction results are more in line with human biological rules.
population mortalityBayesian hierarchical modeldata cloninglongevity risk
万方数据
维普
