Abstract

This paper presents a two-stage approach to summarizing multiple contrastive viewpoints in opinionated text. In the first stage, we use an unsupervised probabilistic approach to model and extract multiple viewpoints in text. We experiment with a variety of lexical and syntactic features, yielding significant performance gains over bag-of-words feature sets. In the second stage, we introduce Comparative LexRank, a novel random walk formulation to score sentences and pairs of sentences from opposite viewpoints based on both their representativeness of the collection as well as their contrastiveness with each other. Experimental results show that the proposed approach can generate informative summaries of viewpoints in opinionated text.

Original languageEnglish (US)
StatePublished - Oct 2010
EventConference on Empirical Methods in Natural Language Processing, EMNLP 2010 - Cambridge, MA, United States
Duration: Oct 9 2010Oct 11 2010

Other

OtherConference on Empirical Methods in Natural Language Processing, EMNLP 2010
CountryUnited States
CityCambridge, MA
Period10/9/1010/11/10

Fingerprint

Syntactics
Experiments

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Science Applications
  • Information Systems

Cite this

Paul, M. J., Zhai, C. X., & Girju, R. (2010). Summarizing contrastive viewpoints in opinionated text. Paper presented at Conference on Empirical Methods in Natural Language Processing, EMNLP 2010, Cambridge, MA, United States.

Summarizing contrastive viewpoints in opinionated text. / Paul, Michael J.; Zhai, Cheng Xiang; Girju, Roxana.

2010. Paper presented at Conference on Empirical Methods in Natural Language Processing, EMNLP 2010, Cambridge, MA, United States.

Research output: Contribution to conferencePaper

Paul, MJ, Zhai, CX & Girju, R 2010, 'Summarizing contrastive viewpoints in opinionated text', Paper presented at Conference on Empirical Methods in Natural Language Processing, EMNLP 2010, Cambridge, MA, United States, 10/9/10 - 10/11/10.
Paul MJ, Zhai CX, Girju R. Summarizing contrastive viewpoints in opinionated text. 2010. Paper presented at Conference on Empirical Methods in Natural Language Processing, EMNLP 2010, Cambridge, MA, United States.
Paul, Michael J. ; Zhai, Cheng Xiang ; Girju, Roxana. / Summarizing contrastive viewpoints in opinionated text. Paper presented at Conference on Empirical Methods in Natural Language Processing, EMNLP 2010, Cambridge, MA, United States.
@conference{f62b47de5e3a4381b77fb1f9819bb8a3,
title = "Summarizing contrastive viewpoints in opinionated text",
abstract = "This paper presents a two-stage approach to summarizing multiple contrastive viewpoints in opinionated text. In the first stage, we use an unsupervised probabilistic approach to model and extract multiple viewpoints in text. We experiment with a variety of lexical and syntactic features, yielding significant performance gains over bag-of-words feature sets. In the second stage, we introduce Comparative LexRank, a novel random walk formulation to score sentences and pairs of sentences from opposite viewpoints based on both their representativeness of the collection as well as their contrastiveness with each other. Experimental results show that the proposed approach can generate informative summaries of viewpoints in opinionated text.",
author = "Paul, {Michael J.} and Zhai, {Cheng Xiang} and Roxana Girju",
year = "2010",
month = "10",
language = "English (US)",
note = "Conference on Empirical Methods in Natural Language Processing, EMNLP 2010 ; Conference date: 09-10-2010 Through 11-10-2010",

}

TY - CONF

T1 - Summarizing contrastive viewpoints in opinionated text

AU - Paul, Michael J.

AU - Zhai, Cheng Xiang

AU - Girju, Roxana

PY - 2010/10

Y1 - 2010/10

N2 - This paper presents a two-stage approach to summarizing multiple contrastive viewpoints in opinionated text. In the first stage, we use an unsupervised probabilistic approach to model and extract multiple viewpoints in text. We experiment with a variety of lexical and syntactic features, yielding significant performance gains over bag-of-words feature sets. In the second stage, we introduce Comparative LexRank, a novel random walk formulation to score sentences and pairs of sentences from opposite viewpoints based on both their representativeness of the collection as well as their contrastiveness with each other. Experimental results show that the proposed approach can generate informative summaries of viewpoints in opinionated text.

AB - This paper presents a two-stage approach to summarizing multiple contrastive viewpoints in opinionated text. In the first stage, we use an unsupervised probabilistic approach to model and extract multiple viewpoints in text. We experiment with a variety of lexical and syntactic features, yielding significant performance gains over bag-of-words feature sets. In the second stage, we introduce Comparative LexRank, a novel random walk formulation to score sentences and pairs of sentences from opposite viewpoints based on both their representativeness of the collection as well as their contrastiveness with each other. Experimental results show that the proposed approach can generate informative summaries of viewpoints in opinionated text.

UR - http://www.scopus.com/inward/record.url?scp=80053271540&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053271540&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:80053271540

ER -