TABLE II:

Test AUROC for vanilla supervised learning. The best results are shown in bold and the second best are shown in underlined.

Methods	Datasets
	CG	CA	DS	AD	CB	BL	IO	IC
LR	0.720	0.836	0.557	0.851	0.748	0.801	0.769	0.860
XGBoost	0.726	0.895	0.587	0.912	0.892	0.821	0.758	0.925
MLP	0.643	0.832	0.568	0.904	0.613	0.832	0.779	0.893
SNN	0.641	0.880	0.540	0.902	0.621	0.834	0.794	0.892
TabNet	0.585	0.800	0.478	0.904	0.680	0.819	0.742	0.896
DCN	0.739	0.870	0.674	0.913	0.848	0.840	0.768	0.915
AutoInt	0.744	0.866	0.672	0.913	0.808	0.844	0.762	0.916
TabTrans	0.718	0.860	0.648	0.914	0.855	0.820	0.794	0.882
FT-Trans	0.739	0.859	0.657	0.913	0.862	0.841	0.793	0.915
VIME	0.735	0.852	0.485	0.912	0.769	0.837	0.786	0.908
SCARF	0.733	0.861	0.663	0.911	0.719	0.833	0.758	0.905
TransTab	0.768	0.881	0.643	0.907	0.851	0.845	0.822	0.919
UniTabE-S	0.760	0.930	0.620	0.910	0.850	0.840	0.740	—
MambaTab-D	0.771	0.954	0.643	0.906	0.862	0.852	0.785	0.906
MambaTab-T	0.801	0.963	0.681	0.914	0.896	0.854	0.812	0.920