Table 3.

Performance indices of extreme gradient boosting machine models in predicting hypotension (Defined as a reduction in mean arterial pressure ≥ 20 mmHg from the initial value within 6 h) according to the number of features.

No. of features	AUROC (95% CI)	Accuracy	F1 score	MCC	Brier’s score
5	0.795 (0.759–0.831)	0.691	0.612	0.420	0.159
10	0.813 (0.779–0.848)	0.783	0.642	0.487	0.153
20	0.815 (0.781–0.849)	0.766	0.634	0.467	0.152
30	0.818 (0.785–0.852)	0.756	0.636	0.463	0.154
40	0.820 (0.786–0.853)	0. 787	0.650	0.497	0.155
50	0.823 (0.790–0.856)	0.736	0.642	0.469	0.170
60	0.822 (0.789–0.855)	0.800	0.641	0.505	0.151
70	0.827 (0.794–0.861)	0.792	0.648	0.499	0.148

AUROC area under the receiver operating characteristic curve, CI confidence interval, MCC Matthews correlation coefficient.