Fining active membership functions in fuzzy data mining

Tzung Pei Hong*, Chun Hao Chen, Yu Lung Wu, S. Tseng

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapter

5 Scopus citations

Abstract

This chapter proposes a fuzzy data-mining algorithm for extracting both association rules and membership functions from quantitative transactions. The number of membership functions for each item is not predefined, but can be dynamically adjusted. A GA-based framework for finding membership functions suitable for mining problems is proposed. The encoding of each individual is divided into two parts. The control genes are encoded into bit strings and used to determine whether membership functions are active or not. The parametric genes are encoded into real-number strings to represent membership functions of linguistic terms. The fitness of each set of membership functions is evaluated using the fuzzy-supports of the linguistic terms in the large 1-itemsets and the suitability of the derived membership functions. The suitability of membership functions considers overlap, coverage and usage factors.

Original languageEnglish
Title of host publicationData Mining
Subtitle of host publicationFoundations and Practice
Pages179-196
Number of pages18
DOIs
StatePublished - 15 Sep 2008

Publication series

NameStudies in Computational Intelligence
Volume118
ISSN (Print)1860-949X

Fingerprint Dive into the research topics of 'Fining active membership functions in fuzzy data mining'. Together they form a unique fingerprint.

  • Cite this

    Hong, T. P., Chen, C. H., Wu, Y. L., & Tseng, S. (2008). Fining active membership functions in fuzzy data mining. In Data Mining: Foundations and Practice (pp. 179-196). (Studies in Computational Intelligence; Vol. 118). https://doi.org/10.1007/978-3-540-78488-3_11