TreeLex / README.md
de-francophones's picture
Update README.md
005d14a verified
metadata
language:
  - fra
configs:
  - config_name: corpusDiCo
    data_files:
      - split: train
        path: treelex_adjs.csv
    sep: ;
  - config_name: corpusDiCo_Hausmann
    data_files:
      - split: train
        path: treelex_verbs.csv
    sep: ;

Description

TreeLex is a syntactic lexicon for French automatically extracted from a French Treebank. It contains syntactic frames of verbs and adjectives present in the corpus. Every entry consists of a lemma, all syntactic realisations (frames) found in the corpus and its frequency in the Treebank.

  • verbs: The lexicon contains 3229 entries corresponding to 1912 verbs (lemmas) present in the corpus. Automatic frame extraction has been based on syntactic annotations present in the Treebank but optional realisations of specific frames have been added manually by Anne Abeillé.
  • adjectives: The lexicon contains 2506 entries corresponding to 2166 adjectives (lemmas) present in the corpus. Automatic frame extraction has been guided by a linguistic study of adjectival constructions in order to complete annotations missing in the corpus.

Citations

Kupść, A., Abeillé A. Growing TreeLex. In Gelbukh, A. (ed.), 9th International Conference (CICLing), Haifa (Israel), February 2008, p. 28-39 (Lecture Notes in Computational Linguistics, 4919)
Kupść, A. Adjectives in TreeLex. In Kłopotek, M., Przepiórkowski, A., Wierzchoń, S., Trojanowski, K. (ed.), 16th International Conference Intelligent Information Systems, Zakopane (Poland), 16-18 juin 2008, Academic Publishing House EXIT, p. 287-296