Estimation of Multiple Logistic Model by Using Empirical Bayes Weights and Comparing it with the Classical Method with Application

In this research, a solution of heteroscedastisity of the random error variance is found in the Multiple Logistic Model when the response variable (dependent) variable is Qualitative by using the method of weighted least squares (WLS) to estimate the parameters of the Multiple Logistic Model, which depends on the weights estimated Empirical Bayes method and compared with the Classical method through some statistical criteria (Mean Square Error (MSE), Coefficient of determination and the F test), so as to obtain the best possible estimate of the parameters of this model through practical application deals with the study of the relationship between the number of recovered patients of severe acute renal concentrations of different types of medicines and are (Ciprodar) and (Garamycin) was given to them in the Republican Hospital / Erbil, and through the design of a program language MATLAB to calculate the weights of Bayes and depend on the statistical package SPSS in the procedures of regression analysis.