A Thesis Submitted to
Council of The Administration and Economics / Karbala University
as Partial fulfillment of the Requirements for the Degree of Master
of Science in Statistics

Researcher
Adnan Fadhel Touma
Supervised By
Assistant Prof. Dr.Shrooq Abd-ALRida Saeed

Abstract
Consider the negative binomial regression model one of the count models , that used to represent many phenomena and cases , these can not be expressed by ordinary models , necessarily we must use a count models generally and negative binomial regression spetially to represent these phenomena and cases , as well as lack of the researches and modern studies for these threads particularly health topics . Have been estimated the parameters of negative binomial regression model by four different methods : [ maximum likelihood estimation (MLE) ,iteratively re-weighted least square (IRLS) , weighted likelihood estimation (WLE) , weighted least square (WLS) . The goal of the study to reach the best method for estimation , to satisfy this where take a random sample (257) patients are newborn suffering from congenital anomalies , they registered in the department of health Babylon . by using statistical programs (stata , spss , minitab ) to estimate the parameters of the models and determine the best method.
The obtained results indicates that the method of (IRLS)is the best , since it shown smallest mean square error ( MSE) and the highest R2