Table 7 Results on IMDb-MLH-mb (1000 features) and positives selection AL3A
From: Co-MLHAN: contrastive learning for multilayer heterogeneous attributed networks
Method | Early-stopping | F1 micro | F1 macro | AUC |
|---|---|---|---|---|
Co-MLHAN (EL) | No | 0.7401 ± 0.0089 | 0.7475 ± 0.0084 | 0.8676 ± 0.0052 |
Co-MLHAN-SA (EL) | No | 0.7411 ± 0.0096 | 0.7552 ± 0.0089 | 0.8740 ± 0.0058 |
Co-MLHAN (NL) | No | 0.8810 ± 0.0071 | 0.8755 ± 0.0070 | 0.9566 ± 0.0032 |
Co-MLHAN-SA (NL) | No | 0.8705 ± 0.0063 | 0.8692 ± 0.0063 | 0.9475 ± 0.0034 |
Co-MLHAN (EL) | Yes | 0.7443 ± 0.0057 | 0.7509 ± 0.0058 | 0.8769 ± 0.0029 |
Co-MLHAN-SA (EL) | Yes | 0.7471 ± 0.0071 | 0.7580 ± 0.0070 | 0.8845 ± 0.0047 |
Co-MLHAN (NL) | Yes | 0.8707 ± 0.0126 | 0.8661 ± 0.0113 | 0.9532 ± 0.0069 |
Co-MLHAN-SA (NL) | Yes | 0.8694 ± 0.0059 | 0.8672 ± 0.0060 | 0.9542 ± 0.0021 |