Graph regularized non-negative matrix factorization with $$L_{2,1}$$ norm regularization terms for drug–target interactions prediction

Table 8 AUC values of different algorithms under $CV_t$ scenario

Method	NR	GPCR	IC	E
iPALM-DLMF	0.797695 (0.0214)	0.886124 (0.0218)	0.948157 (0.0069)	0.938395 (0.0048)
iPALM-DLMF (without NNDSVD)	0.749641 (0.0249)	0.832083 (0.0200)	0.919673 (0.0066)	0.923762 (0.0084)
iPALM-DLMF ( $\lambda _d$ =0)	0.559851 (0.0288)	0.801559 (0.0261)	0.90617 (0.0084)	0.903253 (0.0084)
iPALM-DLMF ( $\lambda _t$=0)	0.498876 (0.0290)	0.553367 (0.0232)	0.582471 (0.0166)	0.547481 (0.0136)
iPALM-DLMF ( $\lambda _l$=0)	0.72897 (0.0133)	0.828976 (0.0184)	0.914893 (0.0066)	0.921384 (0.0054)
PALM-DLMF	0.735579 (0.0248)	0.828103 (0.0175)	0.902373 (0.0009)	0.928485 (0.0069)

The maximum AUC on each dataset is shown in bold. Standard deviation is shown in parentheses

ISSN: 1471-2105