A comparative study of bayesian models for unsupervised sentiment detection

Chenghua Lin, Yulan He, Richard Everson

Research output: Chapter in Book/Report/Conference proceedingOther chapter contribution

Abstract

This paper presents a comparative study of three closely related Bayesian models for unsupervised document level sentiment classification, namely, the latent sentiment model (LSM), the joint sentiment-topic (JST) model, and the Reverse-JST model. Extensive experiments have been conducted on two corpora, the movie review dataset and the multi-domain sentiment dataset. It has been found that while all the three models achieve either better or comparable performance on these two corpora when compared to the existing unsupervised sentiment classification approaches, both JST and Reverse-JST are able to extract sentiment-oriented topics. In addition, Reverse-JST always performs worse than JST suggesting that the JST model is more appropriate for joint sentiment topic detection.
Original languageEnglish
Title of host publicationProceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning
EditorsMirella Lapata
Place of PublicationStroudsburg, PA (US)
PublisherAssociation for Computational Linguistics
Pages144-152
Number of pages9
ISBN (Print)978-1-932432-83-1
Publication statusPublished - 1 Jan 2010
Event14th Conference on Computational Natural Language Learning, CoNLL'10 - Uppsala, Sweden
Duration: 15 Jul 201016 Jul 2010

Conference

Conference14th Conference on Computational Natural Language Learning, CoNLL'10
CountrySweden
CityUppsala
Period15/07/1016/07/10

Fingerprint

Experiments

Cite this

Lin, C., He, Y., & Everson, R. (2010). A comparative study of bayesian models for unsupervised sentiment detection. In M. Lapata (Ed.), Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning (pp. 144-152). Stroudsburg, PA (US): Association for Computational Linguistics.
Lin, Chenghua ; He, Yulan ; Everson, Richard. / A comparative study of bayesian models for unsupervised sentiment detection. Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning. editor / Mirella Lapata. Stroudsburg, PA (US) : Association for Computational Linguistics, 2010. pp. 144-152
@inbook{d484b25bb9084425863c28d71c4dbc21,
title = "A comparative study of bayesian models for unsupervised sentiment detection",
abstract = "This paper presents a comparative study of three closely related Bayesian models for unsupervised document level sentiment classification, namely, the latent sentiment model (LSM), the joint sentiment-topic (JST) model, and the Reverse-JST model. Extensive experiments have been conducted on two corpora, the movie review dataset and the multi-domain sentiment dataset. It has been found that while all the three models achieve either better or comparable performance on these two corpora when compared to the existing unsupervised sentiment classification approaches, both JST and Reverse-JST are able to extract sentiment-oriented topics. In addition, Reverse-JST always performs worse than JST suggesting that the JST model is more appropriate for joint sentiment topic detection.",
author = "Chenghua Lin and Yulan He and Richard Everson",
year = "2010",
month = "1",
day = "1",
language = "English",
isbn = "978-1-932432-83-1",
pages = "144--152",
editor = "Mirella Lapata",
booktitle = "Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning",
publisher = "Association for Computational Linguistics",

}

Lin, C, He, Y & Everson, R 2010, A comparative study of bayesian models for unsupervised sentiment detection. in M Lapata (ed.), Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning. Association for Computational Linguistics, Stroudsburg, PA (US), pp. 144-152, 14th Conference on Computational Natural Language Learning, CoNLL'10, Uppsala, Sweden, 15/07/10.

A comparative study of bayesian models for unsupervised sentiment detection. / Lin, Chenghua; He, Yulan; Everson, Richard.

Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning. ed. / Mirella Lapata. Stroudsburg, PA (US) : Association for Computational Linguistics, 2010. p. 144-152.

Research output: Chapter in Book/Report/Conference proceedingOther chapter contribution

TY - CHAP

T1 - A comparative study of bayesian models for unsupervised sentiment detection

AU - Lin, Chenghua

AU - He, Yulan

AU - Everson, Richard

PY - 2010/1/1

Y1 - 2010/1/1

N2 - This paper presents a comparative study of three closely related Bayesian models for unsupervised document level sentiment classification, namely, the latent sentiment model (LSM), the joint sentiment-topic (JST) model, and the Reverse-JST model. Extensive experiments have been conducted on two corpora, the movie review dataset and the multi-domain sentiment dataset. It has been found that while all the three models achieve either better or comparable performance on these two corpora when compared to the existing unsupervised sentiment classification approaches, both JST and Reverse-JST are able to extract sentiment-oriented topics. In addition, Reverse-JST always performs worse than JST suggesting that the JST model is more appropriate for joint sentiment topic detection.

AB - This paper presents a comparative study of three closely related Bayesian models for unsupervised document level sentiment classification, namely, the latent sentiment model (LSM), the joint sentiment-topic (JST) model, and the Reverse-JST model. Extensive experiments have been conducted on two corpora, the movie review dataset and the multi-domain sentiment dataset. It has been found that while all the three models achieve either better or comparable performance on these two corpora when compared to the existing unsupervised sentiment classification approaches, both JST and Reverse-JST are able to extract sentiment-oriented topics. In addition, Reverse-JST always performs worse than JST suggesting that the JST model is more appropriate for joint sentiment topic detection.

UR - http://www.scopus.com/inward/record.url?scp=79957604857&partnerID=8YFLogxK

UR - http://dl.acm.org/citation.cfm?id=1870586

M3 - Other chapter contribution

AN - SCOPUS:79957604857

SN - 978-1-932432-83-1

SP - 144

EP - 152

BT - Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning

A2 - Lapata, Mirella

PB - Association for Computational Linguistics

CY - Stroudsburg, PA (US)

ER -

Lin C, He Y, Everson R. A comparative study of bayesian models for unsupervised sentiment detection. In Lapata M, editor, Proceeding : CoNLL '10 proceedings of the fourteenth conference on computational natural language learning. Stroudsburg, PA (US): Association for Computational Linguistics. 2010. p. 144-152