Table 3. Model performance analysis of LGBM + LR double-ensemble model with weight values

wall	wex_age	Accuracy, mean (SD)	Specificity, mean (SD)	Sensitivity, mean (SD)	Balanced accuracy, mean (SD)	AUROC, mean (SD)
1.0	0.0	0.6823(0.0038)	0.682(0.0039)	0.6996(0.0164)	0.6908(0.0081)	0.753(0.0080)
0.9	0.1	0.6866(0.0037)	0.6865(0.0038)	0.6955(0.0166)	0.691(0.0082)	0.7538(0.0079)
0.7	0.3	0.6962(0.0031)	0.6964(0.0033)	0.6861(0.0182)	0.6912(0.0086)	0.7541(0.0076)
0.5	0.5	0.706(0.0027)	0.7064(0.0028)	0.6743(0.0137)	0.6903(0.0064)	0.7519(0.0072)
0.3	0.7	0.7111(0.0029)	0.7118(0.0030)	0.6585(0.0131)	0.6851(0.0056)	0.7461(0.0066)
0.1	0.9	0.7082(0.0036)	0.7091(0.0038)	0.6424(0.0140)	0.6757(0.0056)	0.7355(0.0059)
0.0	1.0	0.7041(0.0041)	0.705(0.0043)	0.6356(0.0120)	0.6703(0.0043)	0.7286(0.0056)

Abbreviations: AUROC, area under receiver of characteristics; LGBM, light gradient-boosting machine; LR, logistic regression.
The bold characters were the best-performance model.