Table 20 Performance of 72-hour mortality prediction by multiple classifiers

Logistic regression

Random forest

mean

0.8356 ±0.0014

0.2277 ±0.0037

0.8517 ±0.0033

0.2282 ±0.0071

first

0.7147 ±0.0030

0.1440 ±0.0031

0.7206 ±0.0054

0.1115 ±0.0025

min, max

0.8374 ±0.0019

0.2395 ±0.0031

0.8590 ±0.0042

0.2558 ±0.0080

min, max, mean

0.8475 ±0.0023

0.2469 ±0.0044

0.8607 ±0.0021

0.2494 ±0.0031

min, max, mean, std

0.8484 ±0.0022

0.2506 ±0.0065

0.8589 ±0.0022

0.2498 ±0.0058

min, max, mean, median, Q1, IQR, kurt

0.8426 ±0.0029

0.2377 ±0.0039

0.8627 ±0.0023

0.2493 ±0.0114

SVM

Decision tree

mean

0.8307 ±0.0031

0.2191 ±0.0080

0.6050 ±0.0164

0.0778 ±0.0066

first

0.6950 ±0.0082

0.1147 ±0.0074

0.5356 ±0.0163

0.0572 ±0.0050

min, max

0.8345 ±0.0031

0.2163 ±0.0041

0.5972 ±0.0112

0.0795 ±0.0052

min, max, mean

0.8372 ±0.0042

0.2158 ±0.0035

0.6040 ±0.0165

0.0835 ±0.0082

min, max, mean, std

0.8331 ±0.0043

0.2196 ±0.0088

0.5895 ±0.0118

0.0752 ±0.0054

min, max, mean, median, Q1, IQR, kurt

0.8377 ±0.0023

0.2275 ±0.0058

0.5952 ±0.0121

0.0777 ±0.0058