Skip to main content

Table 5 Performance comparison of five models

From: Using machine learning methods to predict in-hospital mortality of sepsis patients in the ICU

Model Overall performance Discrimination
Brier score 95% CI AUROC 95% CI Sensitivity 95% CI Specificity 95% CI
LASSO 0.108 0.107–0.109 0.829 0.827–0.831 0.744 0.721–0.767 0.754 0.731–0.777
GBM 0.104 0.102–0.105 0.845 0.837–0.853 0.771 0.750–0.792 0.755 0.722–0.789
RF 0.109 0.108–0.109 0.829 0.823–0.834 0.765 0.756–0.774 0.740 0.719–0.761
LR 0.107 0.105–0.108 0.833 0.830–0.838 0.760 0.740–0.780 0.748 0.724–0.772
SAPS II 0.146 0.142–0.150 0.77 0.760–0.780 0.697 0.668–0.725 0.714 0.695–0.734