Link Property Prediction On Ogbl Ddi
评估指标
Ext. data
Number of params
Test Hits@20
Validation Hits@20
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | Ext. data | Number of params | Test Hits@20 | Validation Hits@20 |
|---|---|---|---|---|
| ensemble-learning-for-graph-neural-networks | No | 10512391 | 0.9777 ± 0.0037 | 0.8965 ± 0.0021 |
| reconsidering-the-performance-of-gae-in-link | No | 13816833 | 0.9443 ± 0.0057 | 0.7979 ± 0.0159 |
| node2vec-scalable-feature-learning-for | No | 645249 | 0.2326 ± 0.0209 | 0.3292 ± 0.0121 |
| network-in-graph-neural-network | No | 1618433 | 0.5770 ± 0.1523 | 0.7323 ± 0.0040 |
| 模型 5 | No | 3761665 | 0.8781 ± 0.0474 | 0.8044 ± 0.0404 |
| 模型 6 | No | 2712931 | 0.7654 ± 0.0459 | 0.6927 ± 0.0054 |
| revisiting-graph-neural-networks-for-link-1 | No | 531138 | 0.3056 ± 0.0386 | 0.2849 ± 0.0269 |
| semi-supervised-classification-with-graph | No | 1421571 | 0.6056 ± 0.0869 | 0.6776 ± 0.0095 |
| gidn-a-lightweight-graph-inception-diffusion | No | 3506691 | 0.9542 ± 0.0000 | 0.8258 ± 0.0000 |
| network-in-graph-neural-network | No | 1487361 | 0.5483 ± 0.1581 | 0.7121 ± 0.0038 |
| 模型 11 | No | 976022023 | 0.9972 ± 0.0004 | 0.9956 ± 0.0001 |
| inductive-representation-learning-on-large | No | 1421057 | 0.5390 ± 0.0474 | 0.6262 ± 0.0037 |
| neural-common-neighbor-with-completion-for | No | 1412098 | 0.8232 ± 0.0610 | 0.7172 ± 0.0025 |
| semi-supervised-classification-with-graph | No | 1289985 | 0.3707 ± 0.0507 | 0.5550 ± 0.0208 |
| memory-associated-differential-learning | No | 1228897 | 0.6781 ± 0.0294 | 0.7010 ± 0.0082 |
| 模型 16 | No | 0 | 0.1773 ± 0.0000 | 0.0947 ± 0.0000 |
| can-gnns-learn-link-heuristics-a-concise | No | 5125250 | 0.9549 ± 0.0073 | 0.9098 ± 0.0294 |
| deepwalk-online-learning-of-social | No | 1543913 | 0.2246 ± 0.0290 | Please tell us |
| distance-enhanced-graph-neural-network-for | No | 3760134 | 0.8239 ± 0.0437 | 0.8206 ± 0.0298 |
| 模型 20 | No | 3761665 | 0.9037 ± 0.0193 | 0.8599 ± 0.0286 |
| 模型 21 | No | 0 | 0.1861 ± 0.0000 | 0.0966 ± 0.0000 |
| adaptive-graph-diffusion-networks-with-hop | No | 3506691 | 0.9538 ± 0.0094 | 0.8943 ± 0.0281 |
| counterfactual-graph-learning-for-link | No | 837635 | 0.8608 ± 0.0198 | 0.8405 ± 0.0284 |
| 模型 24 | No | 2910817 | 0.7704 ± 0.0582 | 0.6928 ± 0.0096 |
| from-graph-low-rank-global-attention-to-2-fwl | No | 1576081 | 0.6230 ± 0.0912 | 0.6675 ± 0.0058 |
| edge-proposal-sets-for-link-prediction | No | 1421057 | 0.7495 ± 0.0317 | 0.6696 ± 0.0198 |
| 模型 27 | No | 10235281 | 0.7385 ± 0.0871 | 0.7225 ± 0.0047 |
| pairwise-learning-for-neural-link-prediction | No | 3497473 | 0.9088 ± 0.0313 | 0.8242 ± 0.0253 |
| path-aware-siamese-graph-neural-network-for | No | 3499009 | 0.9284 ± 0.0047 | 0.8306 ± 0.0134 |
| 模型 30 | No | 1763329 | 0.7672 ± 0.0265 | 0.6713 ± 0.0071 |
| open-graph-benchmark-datasets-for-machine | No | 1224193 | 0.1368 ± 0.0475 | 0.3370 ± 0.0264 |