Skip to main content

Table 4 Performance of the rule-based model vs. baseline models applied on the PNCCC 2012 data

From: Understanding the complexity of sepsis mortality prediction via rule discovery and analysis: a pilot study

Method

AUC

p value

SAPS-I

0.612

\(< 0.01\)

SOFA

0.584

\(< 0.01\)

qSOFA

0.502

\(< 0.01\)

SIRS

0.505

\(< 0.01\)

SVM

0.716

0.014

Random Forest

0.532

\(< 0.01\)

LASSO

0.711

\(< 0.01\)

Ridge regression

0.709

0.021

Logistic regression

0.707

0.017

Rule-based

0.788

–

  1. The third column (p value) shows the significance of difference in AUCs (obtained via the Delong test) of each model against the rule-based model