Intelligence Science and Technology Course, Graduate School of Informatics, Kyoto University

• List of Backups
• View the diff.
• View the diff current.
• View the source.
• Go to Chinese Character-level POS Annotation.
  • 1 (2014-04-30 (Wed) 15:03:03)
  • 2 (2016-03-22 (Tue) 18:36:36)
  • 3 (2016-05-03 (Tue) 20:01:31)
  • 4 (2016-06-08 (Wed) 13:51:13)
  • 5 (2016-06-26 (Sun) 16:04:54)
  • 6 (2016-07-24 (Sun) 10:18:20)
  • 7 (2016-07-24 (Sun) 10:20:55)

---

Chinese Character-level POS annotation †

Description †

Augmented CTB5.0 with character-level part-of-speech annotation.

Each Chinese character is in fact created as a word in origin with complete and independent meaning, it should be treated as the actual minimal morphological unit in Chinese language, and therefore should carry specific part-of-speech. For example, the character “打” (beat) is a verb and the character “破” (broken) is an adjective. A word on the other hand, is either single-character, or a compound formed by single-character words. For example, the verb “打破” (break) can be seen as a compound formed by the two single-character words with the construction “verb + adjective”.

This package includes the following files:

• add_cpos_to_ctb.py: a script to patch the original CTB5.0 with character-level POS annotation.
• cpos_dic.txt: a lexicon used to expand words in the original CTB5.0 character-level POS layer.
• README: definition of each character-level POS tag and usage of the script.

For more details, please refer to the paper "Chinese Morphological Analysis with Character-level POS Tagging" (http://www.aclweb.org/anthology/P14-2042)

Usage †

python add_cpos_to_ctb.py [original CTB5 file] cpos_dic.txt > [reannotated CTB5 file]

"original CTB5 file" corresponds to chapters 1−260 in CTB5.0 (LDC2005T01).

Each line of this file is a segmented and PoS tagged sentence:

• 上海_NR 浦东_NR 开发_NN ...

All sentences should be in utf-8.

Download †

Contact †

• MAIL: shen at nlp.ist.i.kyoto-u.ac.jp

     

Website Policy