Μελέτη μοντέλων θνησιμότητας με βάση το πλαίσιο των γενικευμένων γραμμικών μοντέλων
Study of mortality models based on the framework of generalized linear models
View/ Open
Abstract
In the present thesis, the popular Lee – Carter method for modelling and forecasting mortality will be presented. The basic assumptions on which the Lee – Carter model is based will be described as well as some basic methods for estimating model’s parameters. Then, the theory of generalized linear models will be introduced and Poisson, Negative Binomial, and Binomial models will be described. Next, we will study the Lee – Carter model in the light of the theory of generalized linear models. In the next chapter, the above theory will be applied through the statistical package of R to the mortality data of Greece for the years 1981 – 2019 and the ages 0 – 85. To draw conclusions, we will compare the results of some goodness of fit criteria such as the deviance criterion, Akaike and Schwarz information criteria as well as the comparison of plots of normalized residuals against adjusted death values. In the final step, we will compare the results of current thesis with those of the international literature.