Mortality Forecasting: Machine Learning–Assisted Approach (MLAA)

Andrea Nigri , Sapienza, University of Rome

In the field of mortality, the Lee-Carter (L-C) based approach is the best way to forecast mortality rates. Since the first version of the model in 1992, scholars have developed different versions of it, so we could, therefore, define an ”L-C model family” that includes all developments of it. Nevertheless, the first formulation of 1992, remains the benchmark for comparing the performance of future models. The main aim of this paper is to try to fill a gap between demography methodology and statistical learning field. Indeed using data from Human Mortality Database we will attempt to integrate the L-C model with machine learning approach, in particular by using the Neural Network (NN)

See paper

 Presented in Session 3. Population, Development, & the Environment; Data & Methods; Applied Demography