الفهرس Only 14 pages are availabe for public view
 |  | from | 114 |  |  |
| | | from | 114 | | |
Abstract
Longitudinal data differ from other types of data as we take from every subject more than one observation. It is more powerful than cross sectional data for a fixed number of subjects. This data have several types binary, categorical, count and continuous response. This study will concentrate on longitudinal count data that are type of data in which the observations can take only non negative integer values. So this makes our analysis depends on discrete distribution that used in this case like Poisson distribution and negative binomial distribution. Possibility of having missing data makes all traditional methods give biased and inconsistent estimates. The aim of this thesis is to determine the methods that used in this case and which is better. Review for popular methods that used in the case of missing completely at random and missing at random. This thesis will compare the efficiency of this methods and which one is better in each case, so the simulation will be under different sample size and different rate of missing. These methods will be applied in simulation data in addition to real data to be sure about the results under the Poisson distribution. The relative biased and the relative efficiency will be used as criteria to judge which method is better