This study attempts to model a qualitative endogenous variable whose number exceeds two terms. We address the proper technique, which is the multinomial logit techniquemodel, to measure the impairment in sleep disorder in terms of the IAH. Then, a succinct description of the available data is provided. The main results achieved, applying the multinomial logistic regression, are expected and explained. The two variables (sex of the patient and Clinical snoring), exhibit very high coefficients in exponentials, which demonstrate their importance in distinguishing between the different categories (L, M and S). We propose a procedure for calculating rates of good rankings. In applying this procedure, we get a moderate rate of high good ranking. This positive output can be explained by the construction of the estimated values using the multi-class method in which these values were constructed using the conditional probabilities.
Show abstract
Full Text
