Advertisement

Identifying High-vote Patterns

  • Shichao Zhang
  • Chengqi Zhang
  • Xindong Wu
Part of the Advanced Information and Knowledge Processing book series (AI&KP)

Abstract

While traditional multi-database mining algorithms focus on identifying mono-database-mining-like patterns, in this chapter we develop techniques for identifying novel patterns (high-vote patterns) in multi-databases by analyzing local patterns. The number of patterns forwarded from, say, company branches, may be so large that browsing the pattern set, and finding interesting patterns from it, could be rather difficult for a head office. In particular, it is more difficult to identify which of the forwarded patterns are really useful for a company’s applications. In this chapter we design efficient strategies to search for high-vote patterns from local patterns within the branches of a company. As we stated in Chapter 1, this approach is particularly useful in dual-level applications. Another technique we develop provides a good human interface, by constructing a fuzzy logic controller.

Keywords

Membership Function Association Rule Fuzzy Rule Local Pattern Fuzzy Logic Controller 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag London 2004

Authors and Affiliations

  • Shichao Zhang
    • 1
  • Chengqi Zhang
    • 1
  • Xindong Wu
    • 2
  1. 1.FIT, Universityof Technology SydneyAustralia
  2. 2.Department of Computer ScienceUniversity of VermontUSA

Personalised recommendations