Table 3 The AUCs and AUPRs of compared methods based on databases MDAD and aBioflm under fivefold CV.

From: NMGMDA: a computational model for predicting potential microbe–drug associations based on minimize matrix nuclear norm and graph attention network

Methods

AUC

AUPR

MADA

aBioflm

MDAD

aBioflm

HMDAKATZ

0.90118 ± 0.00101

0.90023 ± 0.00212

0.23271 ± 0.00685

0.30669 ± 0.00771

LAGCN

0.86883 ± 0.00703

0.86413 ± 0.01091

0.35712 ± 0.00514

0.36715 ± 0.00556

HMDA-Pred

0.91776 ± 0.00297

0.91756 ± 0.00402

0.02361 ± 0.00093

0.02848 ± 0.00069

MNNMDA

0.96306 ± 0.00202

0.93156 ± 0.00223

0.18920 ± 0.00561

0.19621 ± 0.00782

GSAMDA

0.94917 ± 0.00053

0.93073 ± 0.01206

0.44363 ± 0.00072

0.45103 ± 0.00513

NMGMDA

0.99468 ± 0.00016

0.99467 ± 0.00017

0.90734 ± 0.00019

0.90330 ± 0.00018

Back to article page