Skip to main content
SpringerLink
Log in

EDU-Capsule: aspect-based sentiment analysis at clause level

  • Regular Paper
  • Published:
Knowledge and Information Systems Aims and scope Submit manuscript

Abstract

Many studies on aspect-based sentiment analysis (ABSA) aim to directly predict aspects and polarities at sentence level. However, it is not rare that a long sentence expresses multiple aspects. In this paper, we propose to study ABSA at EDU-level. Elementary discourse unit (EDU) in rhetorical structure theory is an atomic semantic unit, similar to a clause in a sentence. Through manual annotation of 8,823 EDUs, obtained from the SemEval-2014 Task 4 Restaurant Review dataset, we show that more than 97% of EDUs express at most one aspect. Based on this observation, we propose an EDU-level Capsule network for ABSA. EDU-Capsule learns EDU representations within its sentential context for aspect detection and sentiment prediction. EDU-Capsule outperforms strong baselines in our experiments on two benchmark datasets. Both the EDU-level annotations and EDU-Capsule source code are released to support further studies in this area.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Notes

  1. https://github.com/Ting005/EDU_Capsule.

  2. http://138.197.118.157:8000/segbot/ SegBot segmentation achieves an F1-Score of 92.2% on RST-DT Dataset, reported in its original paper [15].

  3. Note that the notion of capsule in our work follows that of [8, 34], which is fundamentally different from the capsule in CapsNet [45].

  4. https://pytorch.org/.

  5. https://huggingface.co/transformers/model_doc/bert.html.

  6. https://alt.qcri.org/semeval2014/task4/data/uploads/semeval14_absa_annotationguidelines.pdf.

  7. The restaurant-2014 dataset and preprocessed Laptop dataset contain 248 (5.23%) and 32 (1.40%) aspects with conflicting sentiments, respectively (refer to Tables 2 and  4 for more detailed statistics). Some literature [8, 10] refers to sentiment prediction on the resultant positive, negative, and neutral sentiment labels as a 3-way/class classification.

References

  1. Liu B (2012) Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies. Morgan & Claypool Publishers

  2. Wang Y, Li S, Yang J (2018) Toward fast and accurate neural discourse segmentation. In: EMNLP, Brussels, Belgium, pp. 962–967

  3. Wang Y, Huang M, Zhu X, Zhao L (2016) Attention-based LSTM for aspect-level sentiment classification. In: EMNLP, Austin, Texas, USA, pp. 606–615

  4. Xue W, Li T (2018) Aspect based sentiment analysis with gated convolutional networks. In: ACL, Melbourne, Australia, pp. 2514–2523

  5. Li X, Bing L, Lam W, Shi B (2018) Transformation networks for target-oriented sentiment classification. In: ACL, Melbourne, Australia, pp. 946–956

  6. Hu M, Zhao S, Zhang L, Cai K, Su Z, Cheng R, Shen X (2019) CAN: constrained attention networks for multi-aspect sentiment analysis. In: EMNLP-IJCNLP, Hong Kong, China, pp. 4600–4609

  7. Jiang Q, Chen L, Xu R, Ao X, Yang M (2019) A challenge dataset and effective models for aspect-based sentiment analysis. In: EMNLP-IJCNLP, Hong Kong, China, pp. 6279–6284

  8. Wang Y, Sun A, Huang M, Zhu X (2019) Aspect-level sentiment analysis using as-capsules. In: WWW, San Francisco, CA, USA, pp. 2033–2044

  9. Sun C, Huang L, Qiu X (2019) Utilizing BERT for aspect-based sentiment analysis via constructing auxiliary sentence. In: NAACL-HLT, Minneapolis, MN, USA, pp. 380–385

  10. Wu Z, Ong D.C (2021) Context-guided BERT for targeted aspect-based sentiment analysis. In: AAAI, Online, pp. 14094–14102

  11. Mann W.C, Thompson S.A (1986) Rhetorical structure theory: Description and construction of text structures. marina del rey. CA: Information Sciences Institute

  12. Galley M, McKeown K.R, Fosler-Lussier E, Jing H (2003) Discourse segmentation of multi-party conversation. In: ACL, Sapporo, Japan, pp. 562–569

  13. Misra H, Yvon F, Jose J.M, Cappé O (2009) Text segmentation via topic modeling: an analytical study. In: CIKM, Hong Kong, China, pp. 1553–1556

  14. Cardoso P.C.F, Taboada M, Pardo T.A.S (2013) On the contribution of discourse structure to topic segmentation. In: SIGDIAL, Metz, France, pp. 92–96

  15. Li J, Sun A, Joty S.R (2018) Segbot: A generic neural text segmentation model with pointer network. In: IJCAI, Stockholm, Sweden, pp. 4166–4172

  16. Zhou Y (2013) Fine-grained sentiment analysis with discourse structure. PhD thesis, Master’s thesis, Saarland University, Germany

  17. Lazaridou A, Titov I, Sporleder C (2013) A bayesian model for joint unsupervised induction of sentiment, aspect and discourse representations. In: ACL, pp. 1630–1639

  18. Bhatia P, Ji Y, Eisenstein J (2015) Better document-level sentiment analysis from RST discourse parsing. In: EMNLP, Lisbon, Portugal, pp. 2212–2218

  19. Ji Y, Smith N.A (2017) Neural discourse structure for text categorization. In: ACL, Vancouver, Canada, pp. 996–1005

  20. Zhang M, Qian T (2020) Convolution over hierarchical syntactic and lexical graphs for aspect level sentiment analysis. In: EMNLP, Online, pp. 3540–3549

  21. Tang H, Ji D, Li C, Zhou Q (2020) Dependency graph enhanced dual-transformer structure for aspect-based sentiment classification. In: ACL, Online, pp. 6578–6588

  22. Hu M, Liu B (2004) Mining opinion features in customer reviews. In: AAAI, San Jose, California, USA, pp. 755–760

  23. Zhang L, Liu B, Lim S.H, O’Brien-Strain E (2010) Extracting and ranking product features in opinion documents. In: COLING, pp. 1462–1470. Chinese Information Processing Society of China, Beijing, China

  24. Kiritchenko S, Zhu X, Cherry C, Mohammad S.M (2014) Nrc-canada-2014: Detecting aspects and sentiment in customer reviews. In: SemEval@COLING, Dublin, Ireland, pp. 437–442

  25. Zhou X, Wan X, Xiao J (2015) Representation learning for aspect category detection in online reviews. In: AAAI, pp. 417–424. AAAI Press, Austin, Texas, USA

  26. Xue W, Zhou W, Li T, Wang Q (2017) MTNA: A neural multi-task model for aspect category classification and aspect term extraction on restaurant reviews. In: IJCNLP, Taipei, Taiwan, pp. 151–156

  27. Tang D, Qin B, Feng X, Liu T (2016) Effective lstms for target-dependent sentiment classification. In: COLING, Osaka, Japan, pp. 3298–3307

  28. Ma D, Li S, Zhang X, Wang H (2017) Interactive attention networks for aspect-level sentiment classification. In: IJCAI, Melbourne, Australia, pp. 4068–4074

  29. Wang W, Pan S.J, Dahlmeier D, Xiao X (2017) Coupled multi-layer attentions for co-extraction of aspect and opinion terms. In: AAAI, California, USA, pp. 3316–3322

  30. Wang J, Li J, Li S, Kang Y, Zhang M, Si L, Zhou G (2018) Aspect sentiment classification with both word-level and clause-level attention networks. In: IJCAI, Stockholm, Sweden, pp. 4439–4445

  31. Tay Y, Tuan L.A, Hui S.C (2018) Learning to attend via word-aspect associative fusion for aspect-based sentiment analysis. In: AAAI, Louisiana, USA, pp. 5956–5963

  32. Yang J, Yang R, Wang C, Xie J (2018) Multi-entity aspect-based sentiment analysis with context, entity and aspect memory. In: AAAI, pp. 6029–6036. AAAI Press, Louisiana, USA

  33. Bao L, Lambert P, Badia T (2019) Attention and lexicon regularized LSTM for aspect-based sentiment analysis. In: ACL, Florence, Italy, pp. 253–259

  34. Wang Y, Sun A, Han J, Liu Y, Zhu X (2018) Sentiment analysis by capsules. In: WWW, Lyon, France, pp. 1165–1174

  35. Zirn C, Niepert M, Stuckenschmidt H, Strube M (2011) Fine-grained sentiment analysis with structural features. In: IJCNLP, Chiang Mai, Thailand, pp. 336–344

  36. Zhang Z, Singh M.P (2018) Limbic: Author-based sentiment aspect modeling regularized with word embeddings and discourse relations. In: EMNLP, Brussels, Belgium, pp. 3412–3422

  37. Hoogervorst R, Essink E, Jansen W, van den Helder M, Schouten K, Frasincar F, Taboada M (2016) Aspect-based sentiment analysis on the web using rhetorical structure theory. In: ICWE, vol. 9671. Lugano, Switzerland, pp. 317–334

  38. Angelidis S, Lapata M (2018) Multiple instance learning networks for fine-grained sentiment analysis. TACL 6:17–31

    Article  Google Scholar 

  39. Li J, Chiu B, Shang S, Shao L (2020) Neural text segmentation and its application to sentiment analysis. TKDE

  40. Yang Y, Wu B, Li L, Wang S (2020) A joint model for aspect-category sentiment analysis with textgcn and bi-gru. In: DSC, pp. 156–163. IEEE, Hong Kong, China

  41. Li Y, Yang Z, Yin C, Pan X, Cui L, Huang Q, Wei T (2020) A joint model for aspect-category sentiment analysis with shared sentiment prediction layer. In: CCL, pp. 388–400. Springer, Hainan, China

  42. Pontiki M, Galanis D, Pavlopoulos J, Papageorgiou H, Androutsopoulos I, Manandhar S (2014) Semeval-2014 task 4: Aspect based sentiment analysis. In: SemEval@COLING, Dublin, Ireland, pp. 27–35

  43. Li M, Zhang L, Ji H, Radke R.J (2019) Keep meeting summaries on topic: Abstractive multi-modal meeting summarization. In: ACL, Florence, Italy, pp. 2190–2196

  44. Gao Y, Wu C, Li J, Joty S.R, Hoi S.C.H, Xiong C, King I, Lyu M.R (2020) Discern: Discourse-aware entailment reasoning network for conversational machine reading. In: EMNLP, Online, pp. 2439–2449

  45. Sabour S, Frosst N, Hinton G.E (2017) Dynamic routing between capsules. In: NIPS, CA, USA, pp. 3856–3866

  46. Kingma D.P, Ba J (2015) Adam: A method for stochastic optimization. In: ICLR, CA, USA

  47. Pennington J, Socher R, Manning C.D (2014) Glove: Global vectors for word representation. In: EMNLP, Doha, Qatar, pp. 1532–1543

  48. Liu B, Hu M, Cheng J (2005) Opinion observer: analyzing and comparing opinions on the web. In: WWW, Chiba, Japan, pp. 342–351

  49. Kumar A, Saini M, Sharan A (2020) Aspect category detection using statistical and semantic association. Comput. Intell. 36(3):1161–1182

    Article  Google Scholar 

  50. Schouten K, van der Weijde O, Frasincar F, Dekker R (2018) Supervised and unsupervised aspect category detection for sentiment analysis with co-occurrence data. IEEE Trans. Cybern. 48(4):1263–1275

    Article  Google Scholar 

  51. Devlin J, Chang M, Lee K, Toutanova K (2019) BERT: pre-training of deep bidirectional transformers for language understanding. In: NAACL-HLT, Minneapolis, MN, USA, pp. 4171–4186

  52. Carlson L, Okurowski ME, Marcu D (2002) Rst discourse treebank. Linguistic Data Consortium, University of Pennsylvania

  53. Carlson L, Marcu D (2001) Discourse tagging reference manual. ISI Technical Report ISI-TR-545 54(2001), 56

Download references

Acknowledgements

This work was partly supported by the National Key RD Program of China (2020AAA0105200) and the National Science Foundation of China (NSFC No.62106249).

Author information

Authors and Affiliations

  1. School of Computer Science and Engineering, Nanyang Technological University, Singapore, Singapore

    Ting Lin & Aixin Sun

  2. Beijing Academy of Artificial Intelligence, Beijing, China

    Yequan Wang

Authors
  1. Ting Lin
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Aixin Sun
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yequan Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Ting Lin or Yequan Wang.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lin, T., Sun, A. & Wang, Y. EDU-Capsule: aspect-based sentiment analysis at clause level. Knowl Inf Syst 65, 517–541 (2023). https://doi.org/10.1007/s10115-022-01797-z

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10115-022-01797-z

Keywords

Search

Navigation