Skip to main content

Table 5 Experimental results of two variants of RLFDDA

From: RLFDDA: a meta-path based graph representation learning model for drug–disease association prediction

Dataset

type

Acc.

AUC.

Prec.

Recall

F1-score

B-dataset

Attribute

0.7555 ± 0.0060

0.8333 ± 0.0063

0.7495 ± 0.0078

0.7676 ± 0.0079

0.7584 ± 0.0056

Network

0.7823 ± 0.0055

0.8654 ± 0.0044

0.7756 ± 0.0071

0.7945 ± 0.0082

0.7849 ± 0.0054

Aggregated

0.7907 ± 0.0061

0.8728 ± 0.0063

0.7821 ± 0.0084

0.8060 ± 0.0078

0.7938 ± 0.0057

C-dataset

Attribute

0.7482 ± 0.0131

0.8039 ± 0.0154

0.7521 ± 0.0166

0.7413 ± 0.0226

0.7464 ± 0.0138

Network

0.8961 ± 0.0103

0.9592 ± 0.0094

0.9023 ± 0.0113

0.8886 ± 0.0218

0.8952 ± 0.0112

Aggregated

0.9006 ± 0.0121

0.9636 ± 0.0047

0.9035 ± 0.0136

0.8972 ± 0.0222

0.9002 ± 0.0129