Summary

UD Indonesian-CSUI is a conversion from an Indonesian constituency treebank in the Penn Treebank format named Kethu that was also a conversion from a constituency treebank built by Dinakaramani et al. (2015). We named this treebank Indonesian-CSUI, since all the three versions of the treebanks were built at Faculty of Computer Science, Universitas Indonesia.

Introduction

UD Indonesian-CSUI treebank was converted automatically from the Kethu treebank, an Indonesian constituency treebank in the Penn Treebank format. The Kethu treebank itself was converted from a consituency treebank built by Dinakaramani et al. (2015).

Other characteristics of the treebank:

• Genre: news in formal Indonesian (the majority is economic news)
• This treebank consists of 1030 sentences and 28K words. We divide CSUI treebank into testing and training dataset:
  • Testing dataset consists of around 10K words
  • Training dataset consists of around 18K words
• Average sentence length is around 27.4 words per-sentence, which is very high compare to the Indonesian-PUD treebank that has average sentence length of 19.4.

Acknowledgments

• The original constituency treebank was built with manual annotation by Arawinda Dinakaramani, Fam Rashel, Andry Luthfi, and Ruli Manurung at Faculty of Computer Science, Universitas Indonesia in 2015.
• The previous treebank was converted to the Penn Treebank format by Ika Alfina and Jessica Naraiswari Arwidarasti in 2019-2020. This PTB version was named Kethu.
• The Kethu treebank was converted automatically to this UD treebank by Alfina et al. (2020).
• The lemma (LEMMA) and morphological features (FEATS) were generated using Aksara and manually corrected.

References

• Ika Alfina, Indra Budi, and Heru Suhartanto. "Tree Rotations for Dependency Trees: Converting the Head-Directionality of Noun Phrases". In Journal of Computer Science, 2020, Vol 16 No 11.
• M. Yudistira Hanifmuti and Ika Alfina. "Aksara: An Indonesian Morphological Analyzer that Conforms to the UD v2 Annotation Guidelines". In Proceeding of the 2020 International Conference of Asian Language Processing (IALP) in Kuala Lumpur, Malaysia, 4-6 Desember 2020.

Changelog

• 2021-11-15 v2.9
  • Added text_en (translation of each sentence to English, generated using Google Translate)
  • Added features Definite (values: Ind, Def), Mood (values:Ind, Imp), NumType (values: Card, Ord), and Polite (values: Form, Infm)
  • Removed feature Poss (value:Yes)
  • Fixed FEATS columns for various words
  • Changed the annotations for "di mana" (where), "yang" (which), 'apa/apakah' (what, whether, adverb in yes-no questions)
  • Changed the annotations for transition words, such as "sementara itu" (meanwhile), "oleh karena itu" (therefore), etc.
  • Fixed udapi bugs (multi-obj, multi-subj, and so on)
• 2020-11-15 v2.7
  • Initial release in Universal Dependencies.

=== Machine-readable metadata (DO NOT REMOVE!) ================================
Data available since: UD v2.7
License: CC BY-SA 4.0
Includes text: yes
Genre: nonfiction news
Lemmas: automatic with corrections
UPOS: converted with corrections
XPOS: converted with corrections
Features: automatic with corrections
Relations: converted with corrections
Contributors: Alfina, Ika; Arwidarasti, Jessica Naraiswari; Hanifmuti, Muhammad Yudistira; Dinakaramani, Arawinda; Manurung, Ruli; Rashel, Fam; Luthfi, Andry
Contributing: here
Contact: ika.alfina@cs.ui.ac.id
===============================================================================