Table 2 The AUC, C-index, HR &MR recall, and HR recall in training set, validation set and total set.

From: Deep learning model for the prediction of all-cause mortality among long term care people in China: a prospective cohort study

 

Training set

Validation set

Total

No. of patients (deceased)

25,071 (14,203)

17,282 (3362)

42,353 (17,565)

AUC (95% CI)

0.791 (0.785–0.796)

0.722 (0.713–0.732)

0.828 (0.824–0.832)

C-index (95% CI)

0.821 (0.808–0.834)

0.877 (0.853–0.901)

0.806 (0.794–0.819)

HR&MR recall (95% CI)

0.602 (0.598–0.606)

0.567 (0.561–0.574)

0.641 (0.638–0.645)

HR recall (95% CI)

0.587 (0.581–0.593)

0.545 (0.536–0.555)

0.583 (0.578–0.588)

  1. AUC area under the curve, C-index concordance index, HR high risk, MR median risk.
Back to article page