Encyclopedia of Database Systems

2009 Edition

Emerging Pattern Based Classification

  • Guozhu Dong
  • Jinyan Li
Reference work entry
DOI: https://doi.org/10.1007/978-0-387-39940-9_5002



The term “emerging-pattern based classification” refers to any classification algorithm that uses emerging patterns to directly build classifiers or to help build/improve other classifiers.

Key Points

The first approach to consider emerging-pattern based classification is CAEP (classification by aggregating emerging patterns) [2]. The main idea is to aggregate (sum) the discriminating power of many of the emerging patterns contained in a case to be classified. The discriminating power of an emerging pattern is often reflected in the support difference of the pattern in the opposing classes. For each class, the emerging patterns of that class contained in the case are aggregated to form a score; the class with the highest score is deemed to be the class of the case. Score normalization can be used to deal with data/battern imbalance between classes. This classification method can lead to high quality classifiers, comparable or...

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

Recommended Reading

  1. 1.
    Alhammady H. and Ramamohanarao K. Using emerging patterns to construct weighted decision trees. IEEE Trans. Knowl. Data Eng., 18(7):865–876, 2006.CrossRefGoogle Scholar
  2. 2.
    Dong G., Zhang X., Wong L., Li J. CAEP: classification by aggregating emerging patterns. Discov. Sci., 30–42, 1999.Google Scholar
  3. 3.
    Li J., Dong G., Ramamohanarao K., and Wong L. DeEPs: a new instance-based lazy discovery and classification system. Mach. Learn., 54(2):99–124, 2004.zbMATHCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2009

Authors and Affiliations

  • Guozhu Dong
    • 1
  • Jinyan Li
    • 2
  1. 1.Wright State UniversityDaytonUSA
  2. 2.Nanyang Technological UniversitySingaporeSingapore