Table 2.

Performance of various autoMLs and traditional ML algorithms for predicting mRS at discharge.

Sr. No	Models	Training accuracy	Testing accuracy	AUC	Sensitivity	Specificity
1	Auto-Gluon	94.09	88.23	0.95	0.74	0.91
2	MLJAR	logloss 0.22*	84.7	0.85	0.83	0.89
3	Auto-Sklearn	94.03	83.52	0.87	0.73	0.89
4	TPOT	93.69	76.47	0.91	0.44	0.75
5	Decision tree classifier	99.99	74.11	0.83	0.78	0.89
6	H2O	89.93	72.9	NA*	0.53	0.84
7	Logistic regression	89.05	65.88	0.78	0.53	0.77
8	Random forest	68.7	65.88	0.65	0.34	0.67
9	kNN	92.12	64.7	0.69	0.46	0.74
10	SVM	89.49	57.64	0.73	0.42	0.71

The values of AUC, Sensitivity, and Specificity are for the testing set. *Due to the technical limitation of the concerned autoML, the values could not be calculated.