Diagnostic estimation of OSAS using binary mixture logistic regression

Binary (Binomial) Logistic Regression is a statistical model that can be used for classification. Concerning the targeted outcome, if the variance of observations is higher than the variance of expectations, because of overdispersion the success rate of the method in classification goes down. This overdispersion is thought as arising from the unobserved heterogen samples in the data set. In Composite models, the overdispersion is minimized by clustering the data into homogeneous subsets and performing a subset based process. In this study a composite binary logistic regression was used for estimating the sleep apnea. Through this model, snoring signals were classified and with a 98.16% success rate the apnea was diagnosed.