Skip to main content

Table 5 Predictive performance with different feature type

From: Multi-view heterogeneous molecular network representation learning for protein–protein interaction prediction

Feature Type

Acc.

Sen.

Pre.

AUC

AUPR

Inter-view feature

0.7491 ± 0.0090

0.6945 ± 0.0109

0.7797 ± 0.0103

0.8206 ± 0.0080

0.8185 ± 0.0181

Intra-view feature

0.8570 ± 0.0045

0.8130 ± 0.0105

0.8916 ± 0.0099

0.9240 ± 0.0046

0.9238 ± 0.0093

Aggregated feature

0.8655 ± 0.0050

0.8249 ± 0.0085

0.8979 ± 0.0088

0.9301 ± 0.0050

0.9308 ± 0.0045

  1. Best results are bolded