Results
This is the main ranking of the systems by their macro-averaged LAS F1 score.
The ±range corresponds to 95% confidence interval computed with bootstrap resampling.
The [OK]
means all 81 test sets were parsed, otherwise the number in square brackets is the number of test sets with non-zero LAS.
More details are provided at separate pages linked below.
1. Stanford (Stanford) 76.30 ± 0.12 [OK] 2. C2L2 (Ithaca) 75.00 ± 0.12 [OK] 3. IMS (Stuttgart) 74.42 ± 0.13 [OK] 4. HIT-SCIR (Harbin) 72.11 ± 0.14 [OK] 5. LATTICE (Paris) 70.93 ± 0.13 [OK] 6. NAIST SATO (Nara) 70.14 ± 0.13 [OK] 7. Koç University (İstanbul) 69.76 ± 0.13 [OK] 8. ÚFAL – UDPipe 1.2 (Praha) 69.52 ± 0.13 [OK] 9. UParse (Edinburgh) 68.87 ± 0.14 [OK] 10. Orange – Deskiñ (Lannion) 68.61 ± 0.13 [OK] 11. TurkuNLP (Turku) 68.59 ± 0.14 [OK] 12. darc (Tübingen) 68.41 ± 0.13 [OK] 13. BASELINE UDPipe 1.1 (Praha) 68.35 ± 0.14 [OK] 14. MQuni (Sydney) 68.05 ± 0.13 [OK] 15. fbaml (Palo Alto) 67.87 ± 0.13 [OK] 16. LyS-FASTPARSE (A Coruña) 67.81 ± 0.13 [OK] 17. LIMSI-LIPN (Paris) 67.72 ± 0.14 [OK] 18. RACAI (București) 67.71 ± 0.13 [OK] 19. IIT Kharagpur (Kharagpur) 67.61 ± 0.14 [OK] 20. naistCL (Nara) 67.59 ± 0.15 [OK] 21. Wanghao-ftd-SJTU (Shanghai) 66.53 ± 0.14 [OK] 22. UALING (Tucson) 65.24 ± 0.13 [OK] 23. Uppsala (Uppsala) 65.11 ± 0.13 [OK] 24. METU (Ankara) 61.98 ± 0.14 [OK] 25. CLCL (Genève) 61.82 ± 0.13 [OK] 26. Mengest (Shanghai) 61.33 ± 0.14 [79] 27. ParisNLP (Paris) 60.02 ± 0.14 [OK] 28. OpenU NLP Lab (Ra'anana) 56.56 ± 0.14 [OK] 29. TRL (Tokyo) 43.07 ± 0.12 [OK] 30. MetaRomance (Santiago de Compostela) 34.05 ± 0.11 [78] 31. UT (Tartu) 21.10 ± 0.05 [27] 32. ECNU (Shanghai) 3.18 ± 0.02 [36] 33. Wenba-NLU (Wuhan) 0.58 ± 0.01 [46]
Other rankings
- LAS per treebank
- UAS per treebank
- CLAS per treebank
- UPOS tagging
- XPOS tagging
- Morphological featuers
- All morphological tags
- Lemmatization
- Sentence segmentation
- Word segmentation
- Tokenization
- LAS including unofficial runs
- Treebanks ranked by best LAS and CLAS
- System run time
- Systems in a nutshell (responses from the questionnaire)
Outputs of system and evaluation runs are available from http://hdl.handle.net/11234/1-2424.