TABLE 2.

The performance of MVIFMDA with all comparison methods.

Metrics ( $mean\pm std$ )	MDHGI	ABMDA	NIMGSA	NIMCGCN	DANE-MDA	MMGCN	MVIFMDA
AUC	$0.8678\pm 0.0008$	$0.9205\pm 0.0036$	$0.8854\pm 0.0022$	$0.9038\pm 0.0064$	$0.9224\pm 0.0006$	$0.9230\pm 0.0023$	$\mathbf{0.9396}\pm 0.0004$
AUPR	$0.0817\pm 0.0012$	$0.1042\pm 0.0154$	$0.1189\pm 0.0089$	$0.1732\pm 0.0174$	$0.2024\pm 0.0030$	$0.2575\pm 0.0008$	$\mathbf{0.2696}\pm 0.0009$
accuracy	$0.9894\pm 0.0004$	$0.9851\pm 0.0088$	$0.9908\pm 0.0008$	$0.9928\pm 0.0007$	$0.9933\pm 0.0008$	$0.9934\pm 0.0001$	$\mathbf{0.9936}\pm 0.0002$
precision	$0.1428\pm 0.0101$	$0.0690\pm 0.0448$	$0.1680\pm 0.0232$	$0.2569\pm 0.0292$	$0.1559\pm 0.0819$	$0.3338\pm 0.0081$	$\mathbf{0.3494}\pm 0.0118$
recall	$0.2200\pm 0.0135$	$0.1416\pm 0.0932$	$0.1891\pm 0.0156$	$0.2146\pm 0.0203$	$0.1683\pm 0.0861$	$\mathbf{0.2986}\pm 0.0067$	$0.2982\pm 0.0080$
F1	$0.1719\pm 0.0109$	$0.0751\pm 0.0452$	$0.1741\pm 0.0161$	$0.2305\pm 0.0202$	$0.1603\pm 0.0826$	$0.3141\pm 0.0007$	$\mathbf{0.3202}\pm 0.0006$
precision@5%	$0.0504\pm 0.0003$	$0.0553\pm 0.0028$	$0.0500\pm 0.0008$	$0.0589\pm 0.0016$	$0.0689\pm 0.0002$	$0.0695\pm 0.0001$	$\mathbf{0.0701}\pm 0.0002$
precision@10%	$0.0329\pm 0.0001$	$0.0382\pm 0.0009$	$0.0323\pm 0.0004$	$0.0371\pm 0.0007$	$0.0410\pm 0.0001$	$0.0412\pm 0.0001$	$\mathbf{0.0416}\pm 0.00003$
recall @5%	$0.5102\pm 0.0027$	$0.5607\pm 0.0279$	$0.5062\pm 0.0078$	$0.5961\pm 0.0165$	$0.6979\pm 0.0018$	$0.7038\pm 0.0010$	$\mathbf{0.7097}\pm 0.0018$
recall @10%	$0.6525\pm 0.0027$	$0.7579\pm 0.0171$	$0.6412\pm 0.0088$	$0.7370\pm 0.0136$	$0.8150\pm 0.0018$	$0.8179\pm 0.0016$	$\mathbf{0.8261}\pm 0.0006$

The bold values indicate the best values in rows.