Mortality modeling at old-age: a mixture model approach

In this paper, we propose a mixture-based model for mortality modeling above age 70. The proposed model is compared with 4 other widely used models: the Beard, Gompertz, Makeham, and Perks models. Our model captures well the mortality rate's behavior at all ages. We applied the method to a country with high-quality data, Japan, and one with lower data quality, Brazil. In the comparative study for the Japanese population, the model presented a better fit to the data, obtaining an absolute mean percentage error of less than 7%, while the other models presented values greater than 30%