Table 2 Performance metrics for 30-day survival models.

From: Machine learning for predicting device-associated infection and 30-day survival outcomes after invasive device procedure in intensive care unit patients

 

C-index

(95% CI)

IBS

(95% CI)

Time-dependent AUC

(95% CI)

Training dataset

 Cox

0.780 (0.779, 0.780)

0.082 (0.082, 0.082)

0.804 (0.801, 0.806)

 EST

0.906 (0.904, 0.907)

0.066 (0.066, 0.066)

0.941 (0.939, 0.943)

 ST

0.698 (0.690, 0.706)

0.088 (0.087, 0.089)

0.778 (0.777, 0.780)

 GBST

0.798 (0.796, 0.800)

0.082 (0.082, 0.082)

0.825 (0.822, 0.828)

 DeepSurv

0.671 (0.653, 0.690)

0.101 (0.099, 0.102)

0.693 (0.690, 0.696)

Validation dataset

 Cox

0.775 (0.774, 0.776)

0.086 (0.086, 0.086)

0.786 (0.781, 0.791)

 EST

0.777 (0.775, 0.779)

0.086 (0.086, 0.086)

0.795 (0.789, 0.801)

 ST

0.692 (0.684, 0.699)

0.092 (0.092, 0.093)

0.757 (0.753, 0.761)

 GBST

0.738 (0.735, 0.741)

0.090 (0.090, 0.090)

0.751 (0.746, 0.756)

 DeepSurv

0.678 (0.642, 0.714)

0.101 (0.099, 0.102)

0.711 (0.708, 0.713)

Back to article page