Figure 3
From: Personalized survival probabilities for SARS-CoV-2 positive patients by explainable machine learning
Predicted individual discrete and cumulative death probabilities. Weekly discrete and cumulative probabilities of death were predicted for all patients in the test set using data prior to their first positive test. Individual probabilities were summarized by the median, 80 and 20 percentiles for patients who died (red) or survived (green) (a). Predicted cumulative death probabilities were summarized by the median (b) for patients who died before 4 weeks (pink), between 4 and 8 weeks (yellow) and after 8 weeks (blue). Individual examples of predicted cumulative (c) and discrete (d) death probabilities for three patients are depicted indicating the time of death (black dot) or censoring (x).
