Jump to content

User:Pfctdayelise/Using Wikipedia as a resource for computational linguistics

From Wikibooks, open books for an open world

The aim of this book is to outline some areas of research in computational linguistics or natural language processing where Wikipedia, and by extension the other Wikimedia projects, have the potential to be valuable resources. It is not intended to serve as an introduction to either of these fields and does not assume any knowledge of Wikipedia.

Description of Wikimedia projects

[edit | edit source]

Witkionary, Wikinews, Commons, Wikibooks, Wikisource, Wikiquote. Languages. Meta & Commons direct translations of help (etc) pages. Who contributes? growth.

Possibly dv.wp is the smallest project which has a translated logo. Less than 200 articles, probably less than 20 legitimate registered users, Dhivehi language, <300000 speakers, official lang of the Maldives.

Database dumps

[edit | edit source]

Post(pre?)-processing tools

[edit | edit source]

Description of the English Wikipedia

[edit | edit source]
  • License
  • Accessible - dumps
  • Coverage - biased to pop-culture and geek topics (best coverage), wikiprojects
  • Format - MOS - but not reliable
  • FAs, cleanup tags
  • RDRs
[edit | edit source]

Categories

[edit | edit source]

Disambiguation pages

[edit | edit source]

Possible tasks

[edit | edit source]

Word sense disambiguation

[edit | edit source]

Word and phrase translation

[edit | edit source]

Web mining, data mining

[edit | edit source]

Machine translation

[edit | edit source]

Geospatial term disambiguation and named entity recognition

[edit | edit source]

Image analysis (?)

[edit | edit source]

Synonymy, abbreviations (RDRs)

[edit | edit source]