Turkish Wordnet

This repository contains the Turkish Wordnet that is built between 2001-2004 as part of the BalkaNet Project. It consists of three files:

• An XML file for the actual wordnet
• .def and .cfg files to visualise the wordnet with the Visdic tool.

To visualise and/or edit Turkish Wordnet, you can download Visdic from the following page:

http://nlp.fi.muni.cz/projekty/visdic/

In Visdic, open the XML file from Files > Dictionaries and follow Visdic documentation for further actions.

The XML Structure of Turkish Wordnet

The XML file consists of a series of <SYNSET> tags, each containing a single synset and its links to other semantically related synsets.

The <ID> tag is the unique synset identifier that links Turkish Wordnet to all other wordnets built upon Princeton WordNet. The identifier ENG20-07172978-n, for example, indicates that this synset was taken from Princeton Wordnet 2.0, has the Princeton Wordnet 2.0 ID 07172978, and is a noun. Similarly, the identifier BILI-60000008 indicates that this synset is one of the Balkanet-specific con- cepts jointly developed by the six members of the Balkanet Consortium, and has the Balkanet Inter-Lingual Index (BILI) identifier 60000008.

The <POS> tag contains part-of-speech information, where ‘n’ denotes a noun, ‘v’ a verb and ‘a’ an adjective.

The <SYNONYM> tag contains the synset members within <LITERAL> tags, and their automatically assigned sense numbers within <SENSE> tags.

The <ILR> tag is used to link the synset to another synset, where the <TYPE> tag indicates the type of the semantic relation in question.

The <DEF> tag contains a brief definition of the concept, either in English or Turkish, and finally, the <BCS> tag indicates whether the synset belongs to one of the three Base Concept sets defined during the Balkanet project.

A single synset containing the synonymous words bitki and nebat ‘plant’ is shown below:

<SYNSET>
   <ID>ENG171-00012420-n</ID>
   <POS>n</POS>
   <SYNONYM>
   	<LITERAL>
   		bitki
   		<SENSE>1</SENSE>
   	</LITERAL>
   	<LITERAL>
   		nebat
   		<SENSE>1</SENSE>
   	</LITERAL>
   </SYNONYM>
   <ILR>
   	ENG171-00003135-n
   	<TYPE>hypernym</TYPE>
   </ILR>
   <ILR>
   	ENG15-06531331-n
   	<TYPE>holo_member</TYPE>
   </ILR>
   <DEF>Bulunduğu yere kökleriyle tutunup gelişen, döl
   veren ve hayatını tamamladıktan sonra kuruyarak
   varlığı sona eren, yosun, ot, ağaç gibi canlıların
   genel adı.
   </DEF>
   <BCS>1</BCS>
</SYNSET>

Publications

Çetinoğlu, Özlem, Orhan Bilgin, and Kemal Oflazer
Turkish Wordnet
In: ed. by Kemal Oflazer and Murat Saraçlar
Theory and Applications of Natural Language Processing. Springer International Publishing
Chap. 15, pp. 317–336.

@incollection{cetinoglu2018,
 title = {{Turkish} {Wordnet}},
  author={{\c{C}}etino{\u{g}}lu, {\"O}zlem
        and Bilgin, Orhan and Oflazer, Kemal},
 chapter = {15},
 pages = {317--336},
 year = {2018},
 booktitle={Turkish Natural Language Processing},
  editor={Oflazer, Kemal and Sara{\c{c}}lar, Murat},
  isbn={9783319901657},
  series={Theory and Applications of Natural Language Processing},
  publisher={Springer International Publishing}
}

Bilgin, Orhan, Özlem Çetinoğlu, and Kemal Oflazer
Building a WordNet for Turkish
In: Romanian Journal of Information Science and Technology
7.1-2, pp. 163–172

@article{bilgin2004,
  title={Building a {WordNet} for {Turkish}},
  author={Bilgin, Orhan and {\c{C}}etino{\u{g}}lu, {\"O}zlem and Oflazer, Kemal},
  journal={Romanian Journal of Information Science and Technology},
  volume={7},
  number={1-2},
  pages={163--172},
  year={2004},
  publisher={Romanian Academy}
}