Diagnosis and Treatment of the Problems Multicollinearity and Separation in the Logistic Regression Model for Patients with Anemic

Abstract

This research dealt with the subject of study logistic regression model, which is one of nonlinear models Taking character more advanced in the process of statistical analysis, which aims to get the high-level estimates of efficiency. Of the most important problems that appear in this model is the separation between the observations of the dependent variable binary response and Multicollinearity between the explanatory variables. As it has been a real data represented injury anemic which have been obtained from kut hospital artificial kidney department through estimation methods according to the modalities Iterative Maximum Likelihood estimators and Stein Logistic Regression Estimators concerning the treatment Multicollinearity problem and penalized maximum likelihood estimators concerning the treatment separation problem and Adjusted Estimators concerning the treatment of separation problem and Multicollinearity problem and that represent the best estimation methods because it has the mean square error (mse) for the logistic regression model.