CALIBRATING THE LEE-CARTER AND THE POISSON LEE-CARTER MODELS VIA NEURAL NETWORKS

Abstract This paper introduces a neural network (NN) approach for fitting the Lee-Carter (LC) and Poisson model on multiple populations. We develop some NNs that replicate structure of individual LC models allow their joint by simultaneously analysing mortality data all considered The NN architecture is specifically designed to calibrate each using available information instead population-specific subset as in traditional estimation schemes. A large set numerical experiments performed countries Human Mortality Database shows effectiveness our approach. In particular, resulting parameter estimates appear smooth less sensitive random fluctuations often present rates’ data, especially low-population countries. addition, forecasting performance results significantly improved well.