Table 2. AUCs for the datasets and outcomes shown in Table 1.

Dataset	CART	DP-AUC	DP-IG	DP-BAY
pneumonia	0.659 (0.611, 0.707)	0.654 (0.626, 0.683)	0.732 (0.699, 0.765)	0.788 (0.766, 0.811)
sepsis-d	0.669 (0.620, 0.718)	0.746 (0.715, 0.776)	0.766 (0.738, 0.795)	0.757 (0.723, 0.792)
sepsis-s	0.658 (0.629, 0.687)	0.670 (0.644, 0.696)	0.710 (0.688, 0.732)	0.716 (0.691, 0.742)
heart failure-d	0.682 (0.653, 0.710)	0.707 (0.683, 0.730)	0.734 (0.708, 0.761)	0.751 (0.731, 0.770)
heart failure-c	0.653 (0.636, 0.670)	0.644 (0.630, 0.658)	0.699 (0.687, 0.712)	0.712 (0.700, 0.725)
HIT	0.818 (0.771, 0.864)	0.847 (0.811, 0.883)	0.830 (0.790, 0.870)	0.849 (0.811, 0.888)
Alzheimer	0.624 (0.589, 0.658)	0.598 (0.575, 0.621)	0.650 (0.648, 0.692)	0.676 (0.654, 0.698)
Mean	0.680 (0.634, 0.727)	0.693 (0.629, 0.756)	0.734 (0.694, 0.773)	0.748 (0.704, 0.792)

For each method the table gives the mean AUC obtained from 20-fold cross-validation and 95% confidence intervals. The last row gives the mean AUC and 95% confidence intervals over all datasets. Highest mean AUC for each outcome is in bold.