Evolutionary discovery of fuzzy concepts in data

Lewis L.H. Chung, Chun Chung Chan

Research output: Journal article publicationJournal articleAcademic researchpeer-review

1 Citation (Scopus)


Given a set of objects characterized by a number of attributes, hidden patterns can be discovered in them for the grouping of similar objects into clusters. If each of these clusters can be considered as exemplifying a certain concept, then the problem concerned can be referred to as a concept discovery problem. This concept discovery problem can be solved to some extent by existing data clustering techniques. However, they may not be applicable when the concept involved is vague in nature or when the attributes characterizing the objects can be qualitative, quantitative, and fuzzy at the same time. To discover such concepts from objects with such characteristics, we propose a Genetic-Algorithm-based technique. By encoding a specific object grouping in a chromosome and a fitness measure to evaluate the cluster quality, the proposed technique is able to discover meaningful fuzzy clusters and assign membership degrees to objects that do not fully exemplify a certain concept. For evaluation, we tested the proposed technique with simulated and real data and the results are found to be very promising.
Original languageEnglish
Pages (from-to)253-268
Number of pages16
JournalBrain and Mind
Issue number2
Publication statusPublished - 1 Aug 2003


  • Concept discovery
  • Fuzzy clustering
  • Genetic algorithm
  • Linguistic terms
  • Machine learning

ASJC Scopus subject areas

  • Neuroscience(all)


Dive into the research topics of 'Evolutionary discovery of fuzzy concepts in data'. Together they form a unique fingerprint.

Cite this