Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus

研究成果: Conference contribution

1 引用 (Scopus)

抄録

This paper proposes a method to compute cross-lingual semantic similarity between synonym sets. By making use of Princeton Annotated Gloss Corpus as the source of target language statistics, the proposed method exhibited promising results in the experiments: More than 73% of the Princeton WordNet synsets were successfully recovered within the top-5 candidates, given a corresponding set of Japanese WordNet synsets. As the proposed method minimally requires that the input to be seen as an apparently synonymous word set, the method could be extended and the performance would be further improved by incorporating richer information such as textual glosses and/or structural constraints posed by the lexical resources at hand.

元の言語English
ホスト出版物のタイトルGWC 2012: 6th International Global Wordnet Conference, Proceedings
出版者Tribun EU s. r. o.
ページ134-141
ページ数8
ISBN(印刷物)9788026302445
出版物ステータスPublished - 2012
外部発表Yes
イベント6th International Global Wordnet Conference, GWC 2012 - Matsue, Japan
継続期間: 2012 1 92012 1 13

Other

Other6th International Global Wordnet Conference, GWC 2012
Japan
Matsue
期間12/1/912/1/13

Fingerprint

Gloss
Synonyms
WordNet
Experiment
Lexical Resources
Statistics
Semantic Similarity
Language

ASJC Scopus subject areas

  • Language and Linguistics
  • Literature and Literary Theory

これを引用

Hayashi, Y. (2012). Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus. : GWC 2012: 6th International Global Wordnet Conference, Proceedings (pp. 134-141). Tribun EU s. r. o..

Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus. / Hayashi, Yoshihiko.

GWC 2012: 6th International Global Wordnet Conference, Proceedings. Tribun EU s. r. o., 2012. p. 134-141.

研究成果: Conference contribution

Hayashi, Y 2012, Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus. : GWC 2012: 6th International Global Wordnet Conference, Proceedings. Tribun EU s. r. o., pp. 134-141, 6th International Global Wordnet Conference, GWC 2012, Matsue, Japan, 12/1/9.
Hayashi Y. Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus. : GWC 2012: 6th International Global Wordnet Conference, Proceedings. Tribun EU s. r. o. 2012. p. 134-141
Hayashi, Yoshihiko. / Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus. GWC 2012: 6th International Global Wordnet Conference, Proceedings. Tribun EU s. r. o., 2012. pp. 134-141
@inproceedings{8540029a4d314a428792b91f63aac449,
title = "Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus",
abstract = "This paper proposes a method to compute cross-lingual semantic similarity between synonym sets. By making use of Princeton Annotated Gloss Corpus as the source of target language statistics, the proposed method exhibited promising results in the experiments: More than 73{\%} of the Princeton WordNet synsets were successfully recovered within the top-5 candidates, given a corresponding set of Japanese WordNet synsets. As the proposed method minimally requires that the input to be seen as an apparently synonymous word set, the method could be extended and the performance would be further improved by incorporating richer information such as textual glosses and/or structural constraints posed by the lexical resources at hand.",
author = "Yoshihiko Hayashi",
year = "2012",
language = "English",
isbn = "9788026302445",
pages = "134--141",
booktitle = "GWC 2012: 6th International Global Wordnet Conference, Proceedings",
publisher = "Tribun EU s. r. o.",

}

TY - GEN

T1 - Computing cross-lingual synonym set similarity by using princeton annotated gloss corpus

AU - Hayashi, Yoshihiko

PY - 2012

Y1 - 2012

N2 - This paper proposes a method to compute cross-lingual semantic similarity between synonym sets. By making use of Princeton Annotated Gloss Corpus as the source of target language statistics, the proposed method exhibited promising results in the experiments: More than 73% of the Princeton WordNet synsets were successfully recovered within the top-5 candidates, given a corresponding set of Japanese WordNet synsets. As the proposed method minimally requires that the input to be seen as an apparently synonymous word set, the method could be extended and the performance would be further improved by incorporating richer information such as textual glosses and/or structural constraints posed by the lexical resources at hand.

AB - This paper proposes a method to compute cross-lingual semantic similarity between synonym sets. By making use of Princeton Annotated Gloss Corpus as the source of target language statistics, the proposed method exhibited promising results in the experiments: More than 73% of the Princeton WordNet synsets were successfully recovered within the top-5 candidates, given a corresponding set of Japanese WordNet synsets. As the proposed method minimally requires that the input to be seen as an apparently synonymous word set, the method could be extended and the performance would be further improved by incorporating richer information such as textual glosses and/or structural constraints posed by the lexical resources at hand.

UR - http://www.scopus.com/inward/record.url?scp=84904606590&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84904606590&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84904606590

SN - 9788026302445

SP - 134

EP - 141

BT - GWC 2012: 6th International Global Wordnet Conference, Proceedings

PB - Tribun EU s. r. o.

ER -