Table 3 Ablation study on eight datasets.
From: Multi-grained contrastive-learning driven MLPs for node classification
Cora | CiteSeer | PubMed | Photo | Computer | DBLP | Chameleon | Ogbn-arxiv | |
|---|---|---|---|---|---|---|---|---|
no all- | 59.67±0.3 | 59.03±0.4 | 73.79±0.7 | 79.33±0.4 | 74.43±0.5 | 71.03±0.6 | 41.30±0.5 | 55.14±0.4 |
no edge- | 61.54±0.4 | 60.53±0.4 | 74.52±0.5 | 80.04±0.4 | 75.05±0.3 | 71.63±0.4 | 41.73±0.4 | 55.85±0.3 |
no feature- | 83.06±0.6 | 73.62±0.4 | 80.67±0.4 | 93.55±0.5 | 87.48±0.5 | 83.38±0.5 | 52.03±0.5 | 72.78±0.4 |
no node- | 83.31±0.5 | 73.82±0.3 | 80.48±0.4 | 93.33±0.3 | 87.40±0.5 | 83.59±0.6 | 52.40±0.5 | 72.92±0.5 |
cMLPs | 83.96±0.4 | 74.31±0.2 | 81.34±0.3 | 94.01±0.2 | 88.06±0.6 | 84.01±0.4 | 52.36±0.3 | 73.35±0.3 |
