ESTIMATION OF ONE-AND-FIVE DIMENSIONAL SURVIVAL FUNCTIONS FOR CATEGORICAL DATA USING ENTROPY

Life tables are used in many fields in demographic and health research They represent an important indicator of death in society. There are two types of life tables; complete life tables are based on the age at death based on single-age categories and are obtained using a comprehensive survey method. The second type is the abbreviated life tables which are based on the age at death of five-year age groups and are obtained by the sample survey method. In this research, the survival function was estimated for the data obtained from the Central Statistical Organization, social and Economic Survey of the Family in Iraq (IHSES II) using parametric methods (the principle of Maximizing Entropy method (POME), and maximum likelihood method (MLE)), as well as the use of A non-parametric approach, the kernel smoothing method (KS), the compared between the estimation methods using (RMSE) and (MAPE). One of the most important conclusions was the emergence of a preference for the (POME) method for the five-age groups, but in the case of the single-age groups, the (KS) method is the best.