Skip to main content
. 2020 Aug 3;1(6):100092. doi: 10.1016/j.patter.2020.100092

Table 3.

The Performance of 12-Variable Models for Identification of Severe COVID-19 on Admission

LDA Logistic Regression Random Forest Decision Tree SVM XGBoost
AUC macro 0.929 0.917 0.903 0.676 – 0.953
F1 weighted 0.891 0.854 0.848 0.769 0.848 0.896
Accuracy 0.892 0.862 0.862 0.800 0.862 0.892
Sensitivity 0.692 0.538 0.462 0.231 0.462 0.846
Specificity 0.942 0.942 0.962 0.942 0.962 0.904