We define the generalized odd log-logistic normal regression with a dispersion systematic component. We obtain a linear representation, some of its properties, and maximum likelihood estimates. Furthermore, we carry out several simulations for different schemes to evaluate the accuracy of the estimators. The robustness of the new regression model is proved by modeling COVID-19 data. The proposed model explains COVID-19 ICU survival times of the patients in a Brazilian hospital.
Digital Object Identifier (DOI)
S. S da Costa, Nicollas and M. Cordeiro, Gauss
"A New Normal Regression with Medical Applications,"
Applied Mathematics & Information Sciences: Vol. 17:
2, Article 13.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol17/iss2/13