The prediction of ellipses using topic model for japanese colloquial inquiry text

Tomohiko Harada; Yoshikatsu Fujita; Kazuhiko Tsuda

doi:10.1016/j.procs.2013.09.219

The prediction of ellipses using topic model for japanese colloquial inquiry text

Tomohiko Harada, Yoshikatsu Fujita, Kazuhiko Tsuda

Faculty of Liberal Arts

University of Tsukuba

Research output: Contribution to journal › Conference article › peer-review

Abstract

Generally inquiries through Web forms and e-mails are increasing. These inquiry texts usually include many informal expressions use of the colloquial style and many omitted words. An omitted word causes the meaning of a sentence to become ambiguous and makes the reader misread and misunderstand a context. In this paper we propose a method to predict omitted words from context and knowledge using topic information. From the results of evaluation experiment, we have confirmed that some of our methods can predict omitted words at the accuracy rate more than 40% for the expression that we used in the experiment.

Original language	English
Pages (from-to)	1311-1318
Number of pages	8
Journal	Procedia Computer Science
Volume	22
DOIs	https://doi.org/10.1016/j.procs.2013.09.219
State	Published - 2013
Event	17th International Conference in Knowledge Based and Intelligent Information and Engineering Systems, KES 2013 - Kitakyushu, Japan Duration: 9 Sep 2013 → 11 Sep 2013

Keywords

Colloquial expressions
Ellipsis
Gibbs sampling
LDA
Statistical topic models

Access to Document

10.1016/j.procs.2013.09.219

Cite this

@article{93dfc622e0f5427da766cfcfaead6dce,

title = "The prediction of ellipses using topic model for japanese colloquial inquiry text",

abstract = "Generally inquiries through Web forms and e-mails are increasing. These inquiry texts usually include many informal expressions use of the colloquial style and many omitted words. An omitted word causes the meaning of a sentence to become ambiguous and makes the reader misread and misunderstand a context. In this paper we propose a method to predict omitted words from context and knowledge using topic information. From the results of evaluation experiment, we have confirmed that some of our methods can predict omitted words at the accuracy rate more than 40\% for the expression that we used in the experiment.",

keywords = "Colloquial expressions, Ellipsis, Gibbs sampling, LDA, Statistical topic models",

author = "Tomohiko Harada and Yoshikatsu Fujita and Kazuhiko Tsuda",

year = "2013",

doi = "10.1016/j.procs.2013.09.219",

language = "英語",

volume = "22",

pages = "1311--1318",

journal = "Procedia Computer Science",

issn = "1877-0509",

publisher = "Elsevier B.V.",

note = "17th International Conference in Knowledge Based and Intelligent Information and Engineering Systems, KES 2013 ; Conference date: 09-09-2013 Through 11-09-2013",

}

TY - JOUR

T1 - The prediction of ellipses using topic model for japanese colloquial inquiry text

AU - Harada, Tomohiko

AU - Fujita, Yoshikatsu

AU - Tsuda, Kazuhiko

PY - 2013

Y1 - 2013

N2 - Generally inquiries through Web forms and e-mails are increasing. These inquiry texts usually include many informal expressions use of the colloquial style and many omitted words. An omitted word causes the meaning of a sentence to become ambiguous and makes the reader misread and misunderstand a context. In this paper we propose a method to predict omitted words from context and knowledge using topic information. From the results of evaluation experiment, we have confirmed that some of our methods can predict omitted words at the accuracy rate more than 40% for the expression that we used in the experiment.

AB - Generally inquiries through Web forms and e-mails are increasing. These inquiry texts usually include many informal expressions use of the colloquial style and many omitted words. An omitted word causes the meaning of a sentence to become ambiguous and makes the reader misread and misunderstand a context. In this paper we propose a method to predict omitted words from context and knowledge using topic information. From the results of evaluation experiment, we have confirmed that some of our methods can predict omitted words at the accuracy rate more than 40% for the expression that we used in the experiment.

KW - Colloquial expressions

KW - Ellipsis

KW - Gibbs sampling

KW - LDA

KW - Statistical topic models

UR - https://www.scopus.com/pages/publications/84896929871

U2 - 10.1016/j.procs.2013.09.219

DO - 10.1016/j.procs.2013.09.219

M3 - 会議記事

AN - SCOPUS:84896929871

SN - 1877-0509

VL - 22

SP - 1311

EP - 1318

JO - Procedia Computer Science

JF - Procedia Computer Science

T2 - 17th International Conference in Knowledge Based and Intelligent Information and Engineering Systems, KES 2013

Y2 - 9 September 2013 through 11 September 2013

ER -

The prediction of ellipses using topic model for japanese colloquial inquiry text

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this