What should a classifier system learn and how should we measure it?

DOI10.1007/S005000100114MaRDI QIDQ1864378zbMATH OpenOpenAlexFDO

Authors Yanyan Li

Publication date 18 March 2003

Published in Soft Computing (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/s005000100114

zbMATH Keywords

classifier system

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Computing methodologies and applications (68U99)

Review text

We consider the issues of how a classifier system should learn to represent a Boolean function, and how we should measure its progress in doing so. We identify four properties which may be desirable of a representation; that it be complete, accurate, minimal and non-overlapping. We distinguish two categories of learning metric, introduce new metrics and evaluate them. We demonstrate the superiority of population state metrics over performance metrics in two situations, and in the process find evidence of XCS's strong bias against overlapping rules.

Recommendations

• Learning classifier systems: a survey
• scientific article; zbMATH DE number 2087898
• scientific article; zbMATH DE number 2087896
• Analysis and improvement of the genetic discovery component of XCS
• scientific article; zbMATH DE number 2040731

Cited in

(11)

• Ten years of genetic fuzzy systems: Current framework and new trends.
• Towards a theory of strong overgeneral classifiers
• Resource management and scalability of the XCSF learning classifier system
• scientific article; zbMATH DE number 2040733 (Why is no real title available?)
• scientific article; zbMATH DE number 2040737 (Why is no real title available?)
• scientific article; zbMATH DE number 4121476 (Why is no real title available?)
• scientific article; zbMATH DE number 2087895 (Why is no real title available?)
• scientific article; zbMATH DE number 2087896 (Why is no real title available?)
• scientific article; zbMATH DE number 2087898 (Why is no real title available?)
• Classifiers that approximate functions
• Imitation guided learning in learning classifier systems

This page was built for publication: What should a classifier system learn and how should we measure it?

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1864378)