Relation Discovery from Thai News Articles Using Association Rule Mining

  • Nichnan Kittiphattanabawon
  • Thanaruk Theeramunkong
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5477)


Association among news articles is useful information for us to track situation related to events, persons, organizations and other concerned issues as well as to detect inconsistency among news. In this paper, we propose an association-based approach towards mining relations in Thai news articles by exploiting coincident terms. This approach first transforms news documents into term-document representation, applies term weighting techniques and generates association by means of statistics. In the work, either unigram or bigram is used for term representation, term frequency, boolean frequency and their modification with inverse document frequency are alternatively applied for term weighting, and confidence or conviction is in turn selected for association measure. Due to this combination, sixteen possible methods are investigated using approximately 811 Thai news of three categories, i.e., politics, economics, and crime. The ranked relations obtained by each method are compared with evaluation done by human. As the result, the method using bigram, term frequency, and conviction achieves the best performance with a rank-order mismatch of 0.84% on the top-50 mined relations. For the top-300 mined relations, the method with bigram, term frequency with inverse document frequency and conviction performs the best with 6.98% rank-order mismatch.


Association Rule Mining Document Relation Text Mining Thai News Document 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Fukumoto, F., Suzuki, Y.: Topic tracking based on bilingual comparable corpora and semisupervised clustering. ACM TALIP 6(3) (2007)Google Scholar
  2. 2.
    Kumaran, G., Allan, J.: Using names and topics for new event detection. In: Proc. of HLT-EMNLP, pp. 121–128 (2005)Google Scholar
  3. 3.
    Kuo, J., Chen, H.: Multidocument summary generation: Using informative and event words. ACM TALIP 7(1), 1–23 (2008)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Montalvo, S., Martnez, R., Casillas, A., Fresno, V.: Multilingual news clustering: Feature translation vs. identification of cognate named entities. Pattern Recognition Letters 28(16), 2305–2311 (2007)CrossRefGoogle Scholar
  5. 5.
    Kalczynski, P.J., Chou, A.: Temporal document retrieval model for business news archives. IPM 41(3), 635–650 (2005)Google Scholar
  6. 6.
    Mengle, S., Goharian, N., Platt, A.: Discovering relationships among categories using misclassification information. In: Proc. of ACM SAC, pp. 932–937 (2008)Google Scholar
  7. 7.
    Lin, F., Liang, C.: Storyline-based summarization for news topic retrospection. Decision Support Systems 45(3), 473–490 (2008)CrossRefGoogle Scholar
  8. 8.
    Jaruskulchai, C., Kruengkrai, C.: A practical text summarizer by paragraph extraction for thai. In: Proc. of IRAL, pp. 9–16 (2003)Google Scholar
  9. 9.
    Sriphaew, K., Theeramunkong, T.: Quality evaluation for document relation discovery using citation information. IEICE Trans. Inf. Syst. E90-D(8), 1225–1234 (2007)CrossRefGoogle Scholar
  10. 10.
    Larose, D.T.: Discovering Knowledge in Data: an Introduction to Data Mining. John Wiley and Sons, Inc., Hoboken (2005)zbMATHGoogle Scholar
  11. 11.
    Brin, S., Motwani, R., Ullman, J., Tsur, S.: Dynamic itemset counting and implication rules for market basket data. In: Proc. of ACM SIGMOD, pp. 255–264 (1997)Google Scholar
  12. 12.
    Jorge, A., Azevedo, P.: An experiment with association rules and classification: Post-bagging and conviction. In: Hoffmann, A., Motoda, H., Scheffer, T. (eds.) DS 2005. LNCS, vol. 3735, pp. 137–149. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  13. 13.
    David, H.: The Method of Paired Comparisons. Oxford University Press, Oxford (1988)zbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2009

Authors and Affiliations

  • Nichnan Kittiphattanabawon
    • 1
  • Thanaruk Theeramunkong
    • 1
  1. 1.Sirindhorn International Institute of TechnologyThammasat UniversityThailand

Personalised recommendations