SubClassOf triple classification on M-YAGO39K (test)
87.02AccuracyEIKE-PRE-EYE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EIKE-PRE-EYESampling strategy=uniform2024.01 | 87.02 | 86.48 | 87.78 | 87.12 | |
| EIKE-PRE-EYESampling strategy=bernoulli2024.01 | 86.81 | 84.33 | 90.47 | 87.29 | |
| TransEllipsoid*2024.01 | 85.5 | 84.2 | 87.4 | 85.8 | |
| TransFG*2024.01 | 84.7 | 78.7 | 94.1 | 85.7 | |
| TransC*2024.01 | 84.4 | 80.7 | 90.4 | 85.3 | |
| TransR*2024.01 | 80 | 73.9 | 92.9 | 82.3 | |
| TransH*2024.01 | 79.1 | 72.8 | 92.9 | 81.6 | |
| EIKE-UNP-MATSampling strategy=uniform2024.01 | 77.59 | 69.81 | 97.22 | 81.27 | |
| EIKE-UNP-MATSampling strategy=bernoulli2024.01 | 77.55 | 69.99 | 96.46 | 81.12 | |
| EIKE-PRE-MATSampling strategy=uniform2024.01 | 77.55 | 70.19 | 95.79 | 81.01 | |
| EIKE-UNP-EYESampling strategy=uniform2024.01 | 77.3 | 69.38 | 97.73 | 81.15 | |
| EIKE-UNP-EYESampling strategy=bernoulli2024.01 | 77.13 | 69.28 | 97.47 | 80.99 | |
| TransE*2024.01 | 76.9 | 72.3 | 87.2 | 79 | |
| EIKE-PRE-MATSampling strategy=bernoulli2024.01 | 76.87 | 70.27 | 93.18 | 80.12 | |
| TransD*2024.01 | 76.1 | 70.7 | 89 | 78.8 | |
| TransCuboid*2024.01 | 75.5 | 69.1 | 92.4 | 79.1 | |
| HolE*2024.01 | 66.6 | 72.3 | 53.7 | 61.7 | |
| DistMult*2024.01 | 60.7 | 71.7 | 35.5 | 47.7 | |
| ComplEx*2024.01 | 59.8 | 65.6 | 41.4 | 50.7 | |
| Concept2Vec2024.01 | 58.9 | 70.4 | 38.5 | 51.9 |