Alternative variants of the Heligman-Pollard model

This paper proposes three models alternative to the classical three-component, nine-parameters, Heligman-Pollard model. These new models are tested with the original Australian datasets used by Heligman and Pollard, and with mortality data from the Valencia Region for the three-year period 1999–2001. When graduating Australian data, the proposed models do not show a strong improvement over the classical model, to which it already adapted fairly well. On the other side, on the Valencia Region dataset, the new models are able to achieve a much better fit, with an up to 60% decrease in the adopted goodness-of-fit score over the Heligman-Pollard model.