Elamur Research
☰
home / papers / 2011 / Opinion Word Expansion and Target Extraction through Double Propagation

Opinion Word Expansion and Target Extraction through Double Propagation

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" does not allow redistributing the text itself. Original — aclanthology.org/J11-1002.pdf

Summary

Opinion mining relies on two building blocks. The first is a sentiment lexicon that fits the domain. The second is knowing what each opinion refers to, called the target, such as a product feature like battery life. General-purpose lexicons miss domain-specific polarity. Earlier target extractors also depended on Web queries, a known product class, or supervised training. Qiu et al. solve both tasks together with a semi-supervised bootstrapping method they call double propagation. It starts from a small seed set of opinion words. A dependency parser then finds three kinds of syntactic links: opinion word to target, opinion word to opinion word, and target to target. Using these links, the method alternates between collecting new opinion words and new noun targets from the items already known. Extra steps assign polarity to newly found opinion words and remove noisy targets. On a standard product review collection of five data sets, the method outperformed state-of-the-art baselines on both tasks. Its main practical value is that it needs only the seeds and in-domain reviews, with no external resources.

Key points

  • The relations are defined with dependency grammar and labeled OT-Rel (opinion–target), OO-Rel (opinion–opinion) and TT-Rel (target–target). Extracted targets are limited to nouns.
  • Target extraction, averaged over data sets D1–D5, scored precision 0.88, recall 0.83 and F-score 0.86 (Prop-dep). Baseline F-scores were 0.82 for Popescu, 0.76 for Hu, 0.54 for PLSA, 0.54 for CRF-D and 0.43 for CRF. Paired t-tests found the gains significant at the 95% level.
  • An ablation without propagation (noProp-dep) reached the highest precision (0.94) but recall fell to 0.66 (F 0.77). The rules themselves are accurate, and propagation is what drives coverage, at the cost of some added noise.
  • Among the baselines, Popescu requires Web search and a known product class. PLSA tends to surface coarse topics instead of fine-grained targets. The Hu and Popescu scores were taken from their original papers.
  • Planned future work covers pruning methods to raise opinion-word precision and learning syntactic relations automatically through pattern mining.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Natural language processing

Cited by

122 papers of our library cite it (of the 140673 whose bibliographies we read).

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/coli_a_00034

@article{qiu2011opinion,
  title = {{Opinion Word Expansion and Target Extraction through Double Propagation}},
  author = {Qiu, Guang and Liu, Bing and Bu, Jiajun and Chen, Chun},
  journal = {Computational Linguistics},
  year = {2011},
  doi = {10.1162/coli\_a\_00034},
  url = {https://doi.org/10.1162/coli_a_00034}
}

References

29 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Breck, Eric, Yejin Choi, and Claire Cardie. 2007. Identifying expressions of opinion in context. In Proceedings of IJCAI’07, pages 2683–2688. Menlo Park, CA. find
  2. Esuli, Andrea and Fabrizio Sebastiani. 2005. Determining the semantic orientation of terms through gloss classification. In Proceedings of CIKM’05, pages 617–624. New York, NY. find
  3. Finkel, Jenny Rose, Trond Grenager, and Christopher Manning. 2005. Incorporating non-local information into information extraction systems by gibbs sampling. In Proceedings of ACL’05, pages 363–370. Stroudsburg, PA. find
  4. Hatzivassiloglou, Vasileios and Kathleen R. McKeown. 1997. Predicting the semantic orientation of adjectives. In Proceedings of ACL’97, pages 174–181. Stroudsburg, PA. find
  5. Hofmann, Thomas. 1999. Probabilistic latent semantic analysis. In Proceedings of UAI’99, pages 289–296. San Francisco, CA. find
  6. Hu, Mingqing and Bing Liu. 2004. Mining and summarizing customer reviews. In Proceedings of SIGKDD’04, pages 168–177. find
  7. Kaji, Nobuhiro and Masaru Kitsuregawa. 2007. Building lexicon for sentiment analysis from massive collection of html documents. In Proceedings of EMNLP’07, pages 1075–1083. find
  8. Kamps, Jaap, Maarten Marx, Robert J. Mokken, and Maarten de Rijke. 2004. Using Wordnet to measure semantic orientation of adjectives. In Proceedings of LREC’04, pages 1115–1118. find
  9. Kanayama, Hiroshi and Tetsuya Nasukawa. 2006. Fully automatic lexicon expansion for domain-oriented sentiment analysis. find
  10. In Proceedings of EMNLP’06, pages 355–363. find
  11. Kim, Soo-Min and Eduard Hovy. 2004. Determining the sentiment of opinions. In Proceedings of COLING’04, pages 1367–1373. find
  12. Kobayashi, Nozomi, Kentaro Inui, and Yuji Matsumoto. 2007. Extracting aspect-evaluation and aspect-of relations in opinion mining. In Proceedings of EMNLP’07, pages 1065–1074. find
  13. Lafferty, John, Andrew McCallum, and Fernando Pereira. 2001. Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In Proceedings of ICML’01, pages 282–289. find
  14. Liu, Bing. 2006. Web Data Mining: Exploring Hyperlinks, Contents and Usage Data. Springer, Berlin. find
  15. Mei, Qiaozhu, Xu Ling, Matthew Wondra, Hang Su, and ChengXiang Zhai. 2007. Topic sentiment mixture: Modeling facets and opinions in weblogs. In Proceedings of WWW’07, pages 171–180. find
  16. Pang, Bo and Lillian Lee. 2008. Opinion Mining and Sentiment Analysis. Now Publishers Inc., Hanover, MA. find
  17. Pang, Bo, Lillian Lee, and Shivakumar Vaithyanathan. 2002. Thumbs up? Sentiment classification using machine learning techniques. In Proceedings of EMNLP’02, pages 79–86. find
  18. Popescu, Ana-Maria and Oren Etzioni. 2005. Extracting product features and opinions from reviews. In Proceedings of EMNLP’05, pages 339–346. find
  19. Scaffidi, Christopher, Kevin Bierhoff, Eric Chang, Mikhael Felker, Herman Ng, and Chun Jin. 2007. Red opal: Product-feature scoring from reviews. In Proceedings of EC’07, pages 182–191. find
  20. Stoyanov, Veselin and Claire Cardie. 2008. Topic identification for fine-grained opinion analysis. In Proceedings of COLING’08, pages 817–824. find
  21. Takamura, Hiroya, Takashi Inui, and Manabu Okumura. 2005. Extracting semantic orientations of words using spin model. In Proceedings of ACL’05, pages 133–140. find
  22. Takamura, Hiroya, Takashi Inui, and Manabu Okumura. 2007. Extracting semantic orientations of phrases from dictionary. In Proceedings of NAACL HLT’07, pages 292–299. find
  23. Tesniere, Lucien. 1959. Elements de Syntaxe Structurale. Librairie C. Klincksieck, Paris. find
  24. Turney, Peter D. 2002. Thumbs up or thumbs down? Semantic orientation applied to unsupervised classification of reviews. In Proceedings of ACL’02, pages 417–424. find
  25. Turney, Peter D. and Michael L. Littman. 2003. Measuring praise and criticism: Inference of semantic orientation from association. ACM Transactions on Information System, 21(4):315–346. find
  26. Wiebe, Janyce. 2000. Learning subjective adjective from corpora. In Proceedings of AAAI’00, pages 735–740. find
  27. Wiebe, Janyce, Theresa Wilson, Rebecca Bruce, Matthew Bell, and Melanie Martin. 2004. Learning subjective language. Computational Linguistics, 30(3):277–308. find
  28. Wong, Tak-Lam, Wai Lam, and Tik-Shun Wong. 2008. An unsupervised framework for extracting and normalizing product attributes from multiple Web sites. In Proceedings of SIGIR’08, pages 35–42. find
  29. Yu, Hong and Vasileios Hatzivassiloglou. 2003. Towards answering opinion questions: Separating facts from opinions and identifying the polarity of opinion sentences. In Proceedings of EMNLP’03, pages 129–136. find