COMPARISON BETWEEN NORMAL METHODS AND WITH GENETIC ALGORITHM IMPROVED METHODS IN THE ESTIMATION OF THE BINARY LOGISTIC REGRESSION MODEL PARAMETERS.

In this research suggested, one of the most important models of nonlinear regression models extensive use in the modeling of applications statistical, in terms of heart disease which is the binary logistic regression model. Some standard methods have been proposed and employed after modifying them by using the genetic algorithm approach in estimation to suit the estimation of the parameters of this of nonlinear regression models, and then making a comparison between two types of the important estimation methods including the standard estimation methods which included the maximum likelihood method, minimum chi-square method, weighted least squares, bayes method, and improved estimation methods developed which by the researcher which included genetic algorithm method depending on the technique estimates MLE, genetic algorithm method depending on the technique estimates MCSE, genetic algorithm method depending on the technique estimates WLSE, genetic algorithm method depending on the technique estimates BE, to choose the best method of estimation by assuming a number of models during simulation and by using the statistical criteria Mean Squares Error (MSE) for estimators for the purpose of comparing the preference of model parameters estimation methods. Generally, improved estimate methods excellence on the normal methods in estimating parameters, as well the (Wls)method is found to be the best one in the first place one among the standard estimation methods, and (Mcs.GA)method is the best among the important estimation methods for the purpose of estimating the parameters for binary logistic regression model because it has less (MSE) for estimators compared to other methods.