The Analysis Of Bayesian Bootstrap Binary Logistic Regression In Modeling The Recovery Rate Of Covid-19 Patients

This study applied the Bayesian binary logistic regression method and the Bootstrap method to model the recovery rate of COVID-19 patients. The aim is to model the recovery rate of COVID-19 patients in order to identify the symptoms and factors affecting the recovery rate of COVID-19 patients. Data obtained from the M. Djamil Hospital, Padang City and Andalas University Hospital on COVID-19 patients in West Sumatra in 2020 were used as the case data. The case data were randomly divided with proportion of 80% training data and 20% testing data. The training data with Bayesian binary logistic regression and perform parameter estimation were later analyzed for testing data using the Bootstrap method. The parameter significance results show that there is one predictor variable that significantly affects the recovery rate of COVID-19 patients, namely patients aged 0 – 59 years. The Bayesian binary logistic regression method used in the modeling has been accurate based on the performance test of the algorithm that has been used with the Bootstrap method. This study proves that the estimated value with Bayesian binary logistic regression is at the 95% Bootstrap confidence interval. The results of the classification model for the recovery rate of COVID-19 patients show good performance by producing high accuracy, sensitivity, and precision values in identifying patients. Therefore, it can be concluded that Bayesian binary logistic regression and the Bootstrap method can be used to model the recovery rate of COVID-19 patients as they produce high classification accurac