الفهرس Only 14 pages are availabe for public view
 |  | from | 165 |  |  |
| | | from | 165 | | |
Abstract
Modeling Mortality has attracted many demographers and mathematicians through history. There are many forms of modeling mortality; mainly tabular form, mathematical forms and a hybrid of them. In mathematical forms, there are two types of models; parametric approach and nonparametric approach. This thesis focuses only on parametric approach of modeling mortality. Parametric approach means that the mortality is represented by parameters. The usual methods of estimating parameters in statistics can be used. In this thesis the used method is Bayesian approach. Heligman Pollard mortality model and Siler mortality model are two of these parametric models. They both model the whole life span; that is why they were chosen to estimate their parameters. Both models have relatively large number of parameters. This makes getting the posterior distribution of their parameters hard analytically. Instead of that, Marcov Chain Monte Carlo (MCMC) method was used- specifically Slice Sampling- to approximate the posterior distribution of the parameters. The thesis aims at fitting both models to different mortality patterns to examine their applicability to different patterns and which one of them fits better. Using World Health Organization (WHO) life tables in 2015, it was showed that Heligman Pollard model is better than Siler mortality model for patterns with mortality hump. However, Siler mortality model is simpler for patterns with small mortality hump or for ages less than 80. Finally both models were applied on Egypt{u2019}s mortality pattern. It was proved that Heligman Pollard model fitted Egypt{u2019}s mortality in 2015 better than Siler model. A time series was obtained for each parameter for both models from 2000-2014. The probability of dying was forecasted for years 2015 and 2020. HP model was able to forecast the probability of dying in 2015 for males better than Siler model. On the other hand, females{u2019} probability of dying in 2015 was forecasted better using Siler model