Advertisement

Journal of Computer Science and Technology

, Volume 26, Issue 1, pp 25–33 | Cite as

Multi-Domain Sentiment Classification with Classifier Combination

  • Shou-Shan Li
  • Chu-Ren Huang
  • Cheng-Qing Zong
Short Paper

Abstract

State-of-the-arts studies on sentiment classification are typically domain-dependent and domain-restricted. In this paper, we aim to reduce domain dependency and improve overall performance simultaneously by proposing an efficient multi-domain sentiment classification algorithm. Our method employs the approach of multiple classifier combination. In this approach, we first train single domain classifiers separately with domain specific data, and then combine the classifiers for the final decision. Our experiments show that this approach performs much better than both single domain classification approach (using the training data individually) and mixed domain classification approach (simply combining all the training data). In particular, classifier combination with weighted sum rule obtains an average error reduction of 27.6% over single domain classification.

Keywords

sentiment classification multiple classifier system multi-domain learning 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Supplementary material

11390_2011_9412_MOESM1_ESM.pdf (76 kb)
(PDF 76.3 kb)

References

  1. [1]
    Pang B, Lee L, Vaithyanathan S. Thumbs up? Sentiment classification using machine learning techniques, In Proc. EMNLP2002, Philadelphia, USA, Jul. 7-12, 2002, pp.79-86.Google Scholar
  2. [2]
    Cui H, Mittal V, Datar M. Comparative experiments on sentiment classification for online product reviews. In Proc. AAAI 2006, Boston, USA, Jul. 16-20, 2006, pp.1265-1270.Google Scholar
  3. [3]
    Kim S, Hovy E. Identifying opinion holders for question answering in opinion texts. In Proc. Workshop on Question Answering in Restricted Domains (AAAI 2005), Pittsburgh, USA, Jul. 9-13, 2005, pp.100-107.Google Scholar
  4. [4]
    Ku L, Liang Y, Chen H. Opinion extraction, summarization and tracking in news and blog corpora. In Proc. the Spring Symposia on Computational Approaches to Analyzing Weblogs (AAAI-CAAW2006), Stanford University, USA, Mar. 27-29, 2006, pp.100-107.Google Scholar
  5. [5]
    Blitzer J, Dredze M, Pereira F. Biographies, Bollywood, boom-boxes and blenders: Domain adaptation for sentiment classification. In Proc. ACL 2007, Prague, Czech, Jun. 23-30, 2007, pp.440-447.Google Scholar
  6. [6]
    Turney P. Thumbs up or thumbs down? Semantic orientation applied to unsupervised classification of reviews. In Proc. ACL 2002, Philadelphia, USA, Jul. 7-12, 2002, pp.417-424.Google Scholar
  7. [7]
    Zagibalov T, Carroll J. Automatic seed word selection for unsupervised sentiment classification of Chinese text. In Proc. COLING 2008, Manchester, UK, Aug. 18-22, 2008, pp.1073-1080.Google Scholar
  8. [8]
    Pang B, Lee L. A sentimental education: Sentiment analysis using subjectivity summarization based on minimum cuts. In Proc. ACL 2004, Barcelona, Spain, Jul. 21-26, 2004, pp.271-278.Google Scholar
  9. [9]
    Riloff E, Patwardhan S, Wiebe J. Feature subsumption for opinion analysis. In Proc. EMNLP2006, Sydney, Australia, Jul. 22-23, 2006, pp.440-448.Google Scholar
  10. [10]
    McDonald R, Hannan K, Neylon T,Wells M, Reynar J. Structured models for fine-to-coarse sentiment analysis. In Proc. ACL 2007, Prague, Czech, Jun. 23-30, 2007, pp.432-439.Google Scholar
  11. [11]
    Aue A, Gamon M. Customizing sentiment classifiers to new domains: A case study. In Proc. RANLP2005, Borovets, Bulgaria, Sept. 21-23, 2005.Google Scholar
  12. [12]
    Li S, Zong C. Multi-domain sentiment classification (short paper). In Proc. ACL 2008, Columbus, USA, Jun. 15-20, 2008, pp.257-260.Google Scholar
  13. [13]
    Dredze M, Crammer K. Online methods for multi-domain learning and adaptation. In Proc. EMNLP2008, Hawaii, USA, Oct. 25-27, 2008, pp.689-697.Google Scholar
  14. [14]
    Daumé III H. Frustratingly easy domain adaptation. In Proc. ACL 2007, Prague, Czech, Jun. 23-30, 2007, pp.256-263.Google Scholar
  15. [15]
    Kittler J, Roli F. Multiple classifier systems. In the First International Workshop on MCS, Cagliari, Italy, Jun. 21-23, 2000.Google Scholar
  16. [16]
    Ranawana R, Palade V. Multi-classifier systems: Review and a roadmap for developers. International Journal of Hybrid Intelligent Systems, 2006, 3(1): 35-61.zbMATHGoogle Scholar
  17. [17]
    Press W, Teukolsky S, Vetterling W, Flannery B. Numerical Recipes in C++: The Art of Scientific Computing, Second Edition. Cambridge University Press, 2002.Google Scholar
  18. [18]
    Vilalta R, Drissi Y. A perspective view and survey of meta-learning. Artificial Intelligence Review, 18(2): 77-95.Google Scholar
  19. [19]
    Roli F, Fumera G. Analysis of linear and order statistics combiners for fusion of imbalanced classifiers. In Proc. MCS2002, Cagliari, Italy, Jun. 24-26, 2002, pp.252-261.Google Scholar
  20. [20]
    Forman G. An extensive empirical study of feature selection metrics for text classification. The Journal of Machine Learning Research, 3: 1533-7928.Google Scholar
  21. [21]
    Wu T, Lin C, Weng R. Probability estimates for multi-class classification by pairwise coupling. The Journal of Machine Learning Research, 5: 975-1005.Google Scholar
  22. [22]
    Koehn P. Statistical significance tests for machine translation evaluation. In Proc. EMNLP2004, Barcelona, Spain, Jul. 25-26, 2004, pp.388-295.Google Scholar
  23. [23]
    Yang Y, Liu X. A re-examination of text categorization methods. In Proc. SIGIR 1999, Berkeley, USA, Aug. 15-19, 1999, pp.42-49.Google Scholar
  24. [24]
    Li S, Zong C. Classifier combining rules under independence assumptions. In Proc. MCS2007, Prague, Czech, May 23-25, 2007, pp.322-332.Google Scholar

Copyright information

© Springer 2011

Authors and Affiliations

  1. 1.NLP Lab, School of Computer Science and TechnologySoochow UniversitySuzhouChina
  2. 2.Department of Chinese and Bilingual StudiesThe Hong Kong Polytechnic UniversityHong KongChina
  3. 3.National Laboratory of Pattern RecognitionInstitute of Automation, Chinese Academy of SciencesBeijingChina

Personalised recommendations