Feature vector quality and distributional similarity

Maayan Geffet; Ido Dagan

Feature vector quality and distributional similarity

Maayan Geffet
, Ido Dagan

Department of Computer Science and Artificial Intelligence

Hebrew University of Jerusalem

Research output: Contribution to conference › Paper › peer-review

24 Scopus citations

Abstract

We suggest a new goal and evaluation criterion for word similarity measures. The new criterion - meaning-entailing substitutability - fits the needs of semantic-oriented NLP applications and can be evaluated directly (independent of an application) at a good level of human agreement. Motivated by this semantic criterion we analyze the empirical quality of distributional word feature vectors and its impact on word similarity results, proposing an objective measure for evaluating feature vector quality. Finally, a novel feature weighting and selection function is presented, which yields superior feature vectors and better word similarity performance.

Original language	English
State	Published - 2004
Event	20th International Conference on Computational Linguistics, COLING 2004 - Geneva, Switzerland Duration: 23 Aug 2004 → 27 Aug 2004

Conference

Conference	20th International Conference on Computational Linguistics, COLING 2004
Country/Territory	Switzerland
City	Geneva
Period	23/08/04 → 27/08/04

Bibliographical note

Publisher Copyright:
© 2004 COLING 2004 - Proceedings of the 20th International Conference on Computational Linguistics. All rights reserved.

Cite this

@conference{a2e0049183ec4f58872544b19140fff7,

title = "Feature vector quality and distributional similarity",

abstract = "We suggest a new goal and evaluation criterion for word similarity measures. The new criterion - meaning-entailing substitutability - fits the needs of semantic-oriented NLP applications and can be evaluated directly (independent of an application) at a good level of human agreement. Motivated by this semantic criterion we analyze the empirical quality of distributional word feature vectors and its impact on word similarity results, proposing an objective measure for evaluating feature vector quality. Finally, a novel feature weighting and selection function is presented, which yields superior feature vectors and better word similarity performance.",

author = "Maayan Geffet and Ido Dagan",

note = "Publisher Copyright: {\textcopyright} 2004 COLING 2004 - Proceedings of the 20th International Conference on Computational Linguistics. All rights reserved.; 20th International Conference on Computational Linguistics, COLING 2004 ; Conference date: 23-08-2004 Through 27-08-2004",

year = "2004",

language = "אנגלית",

}

TY - CONF

T1 - Feature vector quality and distributional similarity

AU - Geffet, Maayan

AU - Dagan, Ido

PY - 2004

Y1 - 2004

N2 - We suggest a new goal and evaluation criterion for word similarity measures. The new criterion - meaning-entailing substitutability - fits the needs of semantic-oriented NLP applications and can be evaluated directly (independent of an application) at a good level of human agreement. Motivated by this semantic criterion we analyze the empirical quality of distributional word feature vectors and its impact on word similarity results, proposing an objective measure for evaluating feature vector quality. Finally, a novel feature weighting and selection function is presented, which yields superior feature vectors and better word similarity performance.

AB - We suggest a new goal and evaluation criterion for word similarity measures. The new criterion - meaning-entailing substitutability - fits the needs of semantic-oriented NLP applications and can be evaluated directly (independent of an application) at a good level of human agreement. Motivated by this semantic criterion we analyze the empirical quality of distributional word feature vectors and its impact on word similarity results, proposing an objective measure for evaluating feature vector quality. Finally, a novel feature weighting and selection function is presented, which yields superior feature vectors and better word similarity performance.

UR - https://www.scopus.com/pages/publications/80053342630

M3 - ???researchoutput.researchoutputtypes.contributiontoconference.paper???

AN - SCOPUS:80053342630

T2 - 20th International Conference on Computational Linguistics, COLING 2004

Y2 - 23 August 2004 through 27 August 2004

ER -

Feature vector quality and distributional similarity

Abstract

Conference

Bibliographical note

Other files and links

Fingerprint

Cite this