Improve the Fitment Quality of the Binary Response Model With Practical Application to those Infected With the Covid-19 Epidemic

Abstract

The study of binary dependent variables is considered one of the important processes because of increasing phenomena that are described by this process. Therefore, the Binary Response model is considered one of the important methods that represent this type of phenomena. Also, the process of choosing independent variables, which affect the Binary Independent variables is considered very necessary. The study included the use of two approaches, the first is the experimental approach (simulation), and the second is the practical application to those infected with the Corona epidemic. Three methods were used to choose the best   binary Response model, which are the forward method, backward  method and the proposed method (factor analysis method) with testing the quality of the model fit after applying each method, by using the Deviation (D) test and Hossmer-Lemshow (H&L) test. The comparison of the final results based on three criteria; Maximum Likelihood Ratio (MLR), Akiake information criteria (AIC) and Bayesian information criteria (BIC). The final results show that the factors resulted from the proposed method (Factor Analysis) have an ability to decrease the MLR better than any group of variables chosen by other methods.  Consequently, AIC and BIC criteria, which are based on MLR gave preferences for the factors that resulted from the proposed method (Factor Analysis) better than the other two methods (forward and backward). The results also showed that the greater the number of independent variables that have a significant effect on the model, the more the factors gave better results according to the criteria used. The results showed that the two independent variables, age of the patient and smoking, have the most impact on the lives of those infected with the Corona epidemic.