Skip to main content
Fig. 3 | BMC Public Health

Fig. 3

From: Predictive determinants of overall survival among re-infected COVID-19 patients using the elastic-net regularized Cox proportional hazards model: a machine-learning algorithm

Fig. 3

Left: The partial likelihood deviance of five-fold cross validation including lower and upper standard deviations (SDs) as a function of log (lambda) for the dataset of the re-infected COVID-19 patients. The dashed and dotted vertical lines demonstrate the lambda values with a minimal deviance (log λ=-3.2226) and the largest lambda value within one SD of the minimal deviance (log λ=-2.8506), respectively. Right: The elastic-net regularized coefficients on the dataset of the re-infected COVID-19 patients are shown as a function of log (lambda)

Back to article page