(Enter summary)
Abstract: We analyse the concept of strong overgeneral rules, the Achilles' heel of traditional
Michigan-style learning classifier systems, using both the traditional strength-based and
newer accuracy-based approaches to rule fitness. We argue that different definitions of
overgenerality are needed to match the goals of the two approaches, present minimal
conditions and environments which will support strong overgeneral rules, demonstrate
their dependence on the reward function, and give some... (Update)
Similar documents (at the sentence level):
40.5%: Rule Fitness and Pathology in Learning - Classifier Systems Tim
(Correct)
Active bibliography (related documents): More All
0.5: Specifying Action Persistence within XCS - Barry (2000)
(Correct)
0.3: A Learning Classifier Systems Bibliography - Kovacs, Lanzi (1999)
(Correct)
0.2: An Indexed Bibliography of Learning Classifier Systems - Alander (1999)
(Correct)
Similar documents based on text: More All
0.5: Strength or Accuracy? - A comparison of two approaches to fitness .. - Kovacs (1999)
(Correct)
0.4: Strength or Accuracy? - Fitness calculation in learning.. - Kovacs (2000)
(Correct)
0.4: Learning Classifier Systems Resources - Kovacs (2002)
(Correct)
BibTeX entry: (Update)
@incollection{ kovacs00towards,
author = "Tim Kovacs",
title = "{Towards a theory of strong overgeneral classifiers}",
booktitle = "{Foundations of Genetic Algorithms Volume 6}",
publisher = "Morgan Kaufmann",
editor = "Worthy Martin and William M. Spears",
pages = "165--184",
year = "2000",
url = "citeseer.ist.psu.edu/437135.html" }
Citations (may not include all citations):
2138
Genetic Algorithms in Search (context) - Goldberg - 1989
614
Reinforcement Learning: An Introduction
- Sutton, Barto - 1998
85
Classifier Fitness Based on Accuracy
- Wilson - 1995
45
ZCS: A Zeroth Level Classifier System
- Wilson - 1994
10
Adding Temporary Memory to ZCS (context) - Cliff, Ross - 1995
10
Empirical Studies of Default Hierarchies and Sequences of Ru.. (context) - Riolo - 1988
9
Default Hierarchy Formation and Memory Exploitation in Learn.. (context) - Smith - 1991
2
Strength or Accuracy (context) - Kovacs - 2000
1
Forthcoming PhD Thesis (context) - Kovacs - 2001
Documents on the same site (ftp://ftp.cs.bham.ac.uk/pub/authors/T.Kovacs/publications.list.html): More
Classifier Systems - Accuracy-Based Fitness Allows (1999)
(Correct)
A Summary of the Attention and Affect Project - Tim Kovacs May
(Correct)
Deletion Schemes for Classifier Systems. A Technical Report. - Kovacs (1999)
(Correct)
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC