A neural network approach for the mortality analysis of multiple populations: a case study on data of the Italian population (Q6593143)

The paper deepens a neural network (NN) approach to jointly model the mortality rates of multiple populations; the empirical results are compared to three classical stochastic mortality models. The Italian mortality data used for the empirical investigation are presented; the index of multiple deprivation (IMD) is employed to categorize the 106 provinces from twenty regions considered in the paper based on their socio-economic indicators. Following that, the NN perspective is presented and the outcomes derived from the NN method are analyzed and compared to those obtained from the model of \N\textit{N. Li} and \textit{R. D. Lee} [Demography 42, No. 3, 575--594 (2005; \url{doi:10.1353/dem.2005.0021})], \Nthe common age effect model of \N\textit{T. Kleinow} [Insur. Math. Econ. 63, 147--152 (2015; Zbl 1348.91233)] \Nand the model of \N\textit{R. Plat} [Insur. Math. Econ. 45, No. 3, 393--404 (2009; Zbl 1231.91227)], \Nrespectively. Technical details are included in the Appendix.