Table 5. Fusion methods on Cora for node classification (micro-F1, metric lies between (0,1) and higher value means better results).
| % Labels | 5% | 10% | 30% | 50% |
|---|---|---|---|---|
| BoW + DeepWalk | 0.74 ± 0.01 | 0.80 ± 0.01 | 0.84 ± 0.00 | 0.86 ± 0.01 |
| Sent2Vec + DeepWalk | 0.76 ± 0.01 | 0.79 ± 0.00 | 0.84 ± 0.01 | 0.85 ± 0.01 |
| TADW - TF-IDF | 0.72 ± 0.02 | 0.80 ± 0.01 | 0.85 ± 0.01 | 0.86 ± 0.01 |
| TADW - Sent2Vec | 0.75 ± 0.01 | 0.80 ± 0.01 | 0.83 ± 0.00 | 0.85 ± 0.00 |
| TADW - Ernie | 0.57 (±0.02) | 0.69 (±0.01) | 0.80 (±0.00) | 0.82 (±0.00) |
| TriDNR | 0.59 ± 0.01 | 0.68 ± 0.00 | 0.75 ± 0.01 | 0.78 ± 0.01 |
| GCN - TF-IDF | 0.80 ± 0.01 | 0.83 ± 0.01 | 0.86 ± 0.01 | 0.87 ± 0.01 |
| GCN - Sent2Vec | 0.77 ± 0.01 | 0.82 ± 0.00 | 0.85 ± 0.01 | 0.87 ± 0.01 |
| GCN - Ernie | 0.60 ± 0.01 | 0.67 ± 0.02 | 0.77 ± 0.01 | 0.81 ± 0.00 |
| GAT - TF-IDF | 0.82 ± 0.02 | 0.84 ± 0.01 | 0.87 ± 0.01 | 0.88 ± 0.00 |
| GAT - Sent2Vec | 0.78 ± 0.00 | 0.81 ± 0.00 | 0.85 ± 0.01 | 0.86 ± 0.00 |
| GAT - Ernie | 0.58 ± 0.02 | 0.62 ± 0.02 | 0.71 ± 0.00 | 0.73 ± 0.00 |
| GraphSAGE - TF-IDF | 0.80 ± 0.01 | 0.84 ± 0.00 | 0.87 ± 0.01 | 0.87 ± 0.01 |
| GraphSAGE - Sent2Vec | 0.75 ± 0.01 | 0.80 ± 0.01 | 0.86 ± 0.01 | 0.88 ± 0.00 |
| GraphSAGE - Ernie | 0.29 ± 0.04 | 0.33 ± 0.05 | 0.34 ± 0.04 | 0.37 ± 0.02 |
| GIC - TF-IDF | 0.74 ± 0.01 | 0.81 ± 0.00 | 0.85 ± 0.00 | 0.88 ± 0.00 |
| GIC - Sent2Vec | 0.66 ± 0.00 | 0.76 ± 0.02 | 0.84 ± 0.00 | 0.86 ± 0.00 |
| GIC - Ernie | 0.34 ± 0.03 | 0.37 ± 0.02 | 0.37 ± 0.01 | 0.38 ± 0.01 |
Note:
The best values with respect to confidence intervals are highlighted in bold.