Skip to main content

Table 2 The AUCs, AUPRs and Accuracy of compared methods based on datasets MDAD and aBiofilm under fivefold CV

From: GSAMDA: a computational model for predicting potential microbe–drug associations based on graph attention network and sparse autoencoder

Methods

AUC

AUPR

Accuracy

MDAD

aBiofilm

MDAD

aBiofilm

MDAD

aBiofilm

HMDAKATZ

0.8712 ± 0.0010

0.8993 ± 0.0021

0.2327 ± 0.0068

0.3066 ± 0.0077

0.9774

0.9796

LAGCN

0.8533 ± 0.0070

0.8641 ± 0.0109

0.3571 ± 0.0051

0.3671 ± 0.0055

0.9413

0.9373

NTSHMDA

0.8483 ± 0.0020

0.8610 ± 0.0022

0.1892 ± 0.0056

0.1962 ± 0.0078

0.9896

0.9882

HMDA-Pred

0.7987 ± 0.0030

0.8053 ± 0.0040

0.0236 ± 0.0009

0.0284 ± 0.0006

0.9794

0.9806

BPNNHMDA

0.8410 ± 0.0320

0.8438 ± 0.0186

0.0319 ± 0.0105

0.0476 ± 0.0067

0.9894

0.9869

GSAMDA

0.9496 ± 0.0005

0.9308 ± 0.0120

0.4436 ± 0.0007

0.4510 ± 0.0051

0.9896

0.9880