Can click patterns across user's query logs predict answers to definition questions?

Resultado de la investigación: Conference contribution

Resumen

In this paper, we examined click patterns produced by users of Yahoo search engine when prompting definition questions. Regularities across these click patterns are then utilized for constructing a large and heterogeneous training corpus for answer ranking. In a nutshell, answers are extracted from clicked web-snippets originating from any class of web-site, including Knowledge Bases (KBs). On the other hand, nonanswers are acquired from redundant pieces of text across web-snippets. The effectiveness of this corpus was assessed via training two state-of-The-Art models, wherewith answers to unseen queries were distinguished. These testing queries were also submitted by search engine users, and their answer candidates were taken from their respective returned web-snippets. This corpus helped both techniques to finish with an accuracy higher than 70%, and to predict over 85% of the answers clicked by users. In particular, our results underline the importance of non-KB training data.

Idioma originalEnglish
Título de la publicación alojadaEACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings
EditorialAssociation for Computational Linguistics (ACL)
Páginas99-108
Número de páginas10
ISBN (versión digital)9781937284190
EstadoPublished - 1 ene 2012
Evento13th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2012 - Avignon, France
Duración: 23 abr 201227 abr 2012

Conference

Conference13th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2012
PaísFrance
CiudadAvignon
Período23/04/1227/04/12

Huella dactilar

Search engines
World Wide Web
Websites
Testing

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Software

Citar esto

Figueroa, A. (2012). Can click patterns across user's query logs predict answers to definition questions? En EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings (pp. 99-108). Association for Computational Linguistics (ACL).
Figueroa, Alejandro. / Can click patterns across user's query logs predict answers to definition questions?. EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings. Association for Computational Linguistics (ACL), 2012. pp. 99-108
@inproceedings{dac6b2ef5b414490a6e9fd806ba50318,
title = "Can click patterns across user's query logs predict answers to definition questions?",
abstract = "In this paper, we examined click patterns produced by users of Yahoo search engine when prompting definition questions. Regularities across these click patterns are then utilized for constructing a large and heterogeneous training corpus for answer ranking. In a nutshell, answers are extracted from clicked web-snippets originating from any class of web-site, including Knowledge Bases (KBs). On the other hand, nonanswers are acquired from redundant pieces of text across web-snippets. The effectiveness of this corpus was assessed via training two state-of-The-Art models, wherewith answers to unseen queries were distinguished. These testing queries were also submitted by search engine users, and their answer candidates were taken from their respective returned web-snippets. This corpus helped both techniques to finish with an accuracy higher than 70{\%}, and to predict over 85{\%} of the answers clicked by users. In particular, our results underline the importance of non-KB training data.",
author = "Alejandro Figueroa",
year = "2012",
month = "1",
day = "1",
language = "English",
pages = "99--108",
booktitle = "EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings",
publisher = "Association for Computational Linguistics (ACL)",

}

Figueroa, A 2012, Can click patterns across user's query logs predict answers to definition questions? En EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings. Association for Computational Linguistics (ACL), pp. 99-108, 13th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2012, Avignon, France, 23/04/12.

Can click patterns across user's query logs predict answers to definition questions? / Figueroa, Alejandro.

EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings. Association for Computational Linguistics (ACL), 2012. p. 99-108.

Resultado de la investigación: Conference contribution

TY - GEN

T1 - Can click patterns across user's query logs predict answers to definition questions?

AU - Figueroa, Alejandro

PY - 2012/1/1

Y1 - 2012/1/1

N2 - In this paper, we examined click patterns produced by users of Yahoo search engine when prompting definition questions. Regularities across these click patterns are then utilized for constructing a large and heterogeneous training corpus for answer ranking. In a nutshell, answers are extracted from clicked web-snippets originating from any class of web-site, including Knowledge Bases (KBs). On the other hand, nonanswers are acquired from redundant pieces of text across web-snippets. The effectiveness of this corpus was assessed via training two state-of-The-Art models, wherewith answers to unseen queries were distinguished. These testing queries were also submitted by search engine users, and their answer candidates were taken from their respective returned web-snippets. This corpus helped both techniques to finish with an accuracy higher than 70%, and to predict over 85% of the answers clicked by users. In particular, our results underline the importance of non-KB training data.

AB - In this paper, we examined click patterns produced by users of Yahoo search engine when prompting definition questions. Regularities across these click patterns are then utilized for constructing a large and heterogeneous training corpus for answer ranking. In a nutshell, answers are extracted from clicked web-snippets originating from any class of web-site, including Knowledge Bases (KBs). On the other hand, nonanswers are acquired from redundant pieces of text across web-snippets. The effectiveness of this corpus was assessed via training two state-of-The-Art models, wherewith answers to unseen queries were distinguished. These testing queries were also submitted by search engine users, and their answer candidates were taken from their respective returned web-snippets. This corpus helped both techniques to finish with an accuracy higher than 70%, and to predict over 85% of the answers clicked by users. In particular, our results underline the importance of non-KB training data.

UR - http://www.scopus.com/inward/record.url?scp=85035309567&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85035309567

SP - 99

EP - 108

BT - EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings

PB - Association for Computational Linguistics (ACL)

ER -

Figueroa A. Can click patterns across user's query logs predict answers to definition questions? En EACL 2012 - 13th Conference of the European Chapter of the Association for Computational Linguistics, Proceedings. Association for Computational Linguistics (ACL). 2012. p. 99-108