Heterogeneous Node Classification On Dblp 1
평가 지표
Macro-F1 (20% training data)
Macro-F1 (60% training data)
Macro-F1 (80% training data)
Micro-F1 (20% training data)
Micro-F1 (80% training data)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | Macro-F1 (20% training data) | Macro-F1 (60% training data) | Macro-F1 (80% training data) | Micro-F1 (20% training data) | Micro-F1 (80% training data) | Paper Title | Repository |
|---|---|---|---|---|---|---|---|
| HAN (sem) | 92.03% | 93.31% | 92.53% | 92.99% | 93.29% | Heterogeneous Graph Attention Network | |
| GAT | 90.97% | 91.84% | 91.73% | 91.96% | 92.55% | Heterogeneous Graph Attention Network | |
| NLAH (ppr) | - | 95.95% | - | - | - | Non-local Attention Learning on Large Heterogeneous Information Networks | |
| GCN | 90.79% | 92.62% | 92.38% | 91.71% | 93.09% | Heterogeneous Graph Attention Network | |
| HDGI-C | 90.94% | - | 91.53% | 91.75% | 92.26% | Heterogeneous Deep Graph Infomax | |
| HAN | 92.24% | 93.70% | 93.08% | 93.11% | 93.99% | Heterogeneous Graph Attention Network | |
| HDGI-A | 89.88% | - | 91.06% | 90.62% | 91.92% | Heterogeneous Deep Graph Infomax | |
| NLAH (ppmi) | - | 95.91% | - | - | - | Non-local Attention Learning on Large Heterogeneous Information Networks | |
| DeepWalk | 77.43% | 85.27% | 84.81% | 79.37% | 86.26% | Heterogeneous Graph Attention Network | |
| NLAH (2ndprox) | - | 96.48% | - | - | - | Non-local Attention Learning on Large Heterogeneous Information Networks | |
| HAN (nd) | 91.17% | 92.69% | 91.80% | 92.05% | 92.69% | Heterogeneous Graph Attention Network | |
| metapath2vec | 90.16% | 92.48% | 91.89% | 91.53% | 92.80% | Heterogeneous Graph Attention Network | |
| HERec | 91.68% | 93.70% | 92.34% | 92.69% | 93.27% | Heterogeneous Graph Attention Network | |
| ESim | 91.64% | 93.39% | 93.44% | 92.73% | 92.53% | Heterogeneous Graph Attention Network |
0 of 14 row(s) selected.