Word Sense Induction on SemEval WSI 2010
18V-Measure (All)NB
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| NBYear=20132016.06 | 18 | 23.7 | 9.9 | 52.9 | 52.5 | 53.5 | 3.42 | |
| SE-WSI-fix-cmpTraining=Jointly learns sense embeddings for 6K words2016.06 | 16.3 | 20.8 | 9.7 | 54.3 | 54.2 | 54.3 | 2.61 | |
| UoYYear=20102016.06 | 15.7 | 20.6 | 8.5 | 49.8 | 38.2 | 66.6 | 11.5 | |
| NMFlibYear=20112016.06 | 11.8 | 13.5 | 9.4 | 45.3 | 42.2 | 49.8 | 4.8 | |
| SE-WSI-fixTraining Data=Out-of-domain2016.06 | 9.8 | 13.5 | 4.3 | 55.1 | 50.7 | 61.6 | 2.5 | |
| SE-WSI-CRPMechanism=Chinese Restaurant Process2016.06 | 5.7 | 7.4 | 3.2 | 55.3 | 49.4 | 63.8 | 2.09 | |
| WE-KmeansTraining=Separate learning2016.06 | 4.6 | 5 | 4.1 | 51.2 | 46.5 | 57.6 | 2.54 | |
| SpectralYear=20142016.06 | 4.5 | 4.6 | 4.2 | 61.5 | 54.5 | 71.6 | 1.87 | |
| CRP-PPMIApproach=Count-based distributional model2016.06 | 2.9 | 3.5 | 2 | 57.7 | 53.3 | 64 | 1.76 |