Nonparametric graduation techniques as a common framework for the description of demographic patterns

Abstract

The graduation of age-specific demographic rates is a subject of special interest in many dis-ciplines as demography, biostatistics, actuarial practice, and social planning. For estimating the unknown age-specific probabilities of the various demographic phenomena, some graduation technique must be applied to the corresponding empirical rates, under the assumption that the true probabilities follow a smooth pattern through age. The classical way for graduating demographic rates is parametric modelling. However, for graduation purposes, nonparametric techniques can also be adapted. This work provides an adaptation, and an evaluation of kernels and Support Vector Machines (SVM) in the context of graduation of demographic rates.