Cross parsers evaluation : a French Treebanks study

Djamé Seddah, Marie Candito and Benoit Crabbé

11th International Conference on Parsing Technology (IWPT 2009)
Paris, France, 7th-9th October, 2009

---

Summary

This paper presents preliminary investigations on the statistical parsing of French by bringing a complete evaluation on French data of the main based probabilistic lexicalized (Charniak, Collins, Chiang) and unlexicalized (Berkeley) parsers designed first on the Penn Treebank. We adapted the parsers on the two existing treebanks of French \cite{abeille:03,schluter:07}. To our knowledge, all the results reported here are state-of-the-art for the constituent parsing of French on every available treebank. Regarding the algorithms, the comparisons show that lexicalized parsing models are outperformed by the unlexicalized Berkeley parser. Regarding the treebanks, we observe that a tag set with a specific feature, has direct influences over evaluation results depending on the parsing model

---

START Conference Manager (V2.56.8 - Rev. 780)