Multinomial Logistic Regression Model for Prioritization of COVID-19 Vaccination in Portoviejo—Ecuador

Abstract:

Currently, the world is in the initial phase of the distribution of the COVID-19 vaccine, and the vaccine is principally available to developed countries, that is mainly administered to older people, especially to health workers at high risk of contracting COVID-19 while the rest of the population are exposed to contagion. A classification method is to classify people with high or low priority for the administration of the vaccine, that is vital importance to curb the spread of infections in the world. Mathematical models can be helped to define the classification while the impact of increased contagion is minimized. A multinomial logistic regression model is proposed to classify subjects, that is based on the values of a set of pbkp_redictor variables. The priority of vaccination is classified in the canton of Portoviejo—Ecuador, the variables are considered: age, sex, number of presented symptoms at the time of registration, cardiovascular, chronic liver, chronic kidney, chronic respiratory, oncological, diabetes, hypertension, tuberculosis, other preexisting disease, exposed days to virus. A stochastic descent gradient algorithm is proposed to minimize an objective function J(θ), that is obtained from the proposed model. The efficiency of the forecasts of the model is compared, that is reproducing accuracy in the estimates. Finally, one goodness-of-fit measure to validate the performance of the model is used, obtaining insignificant estimation error.

Año de publicación:

2022

Keywords:

• Stochastic descent gradient
• Portoviejo—Ecuador
• Multinomial logistic regression model
• COVID-19 Vaccine

Fuente:

scopus