Class imbalance problem in UCS classifier system: fitness adaptation

The class imbalance problem has been said to challenge the performance of concept learning systems. Learning systems tend to be biased towards the majority class, and thus have poor generalization for the minority class instances. We analyze the class imbalance problem in learning classifier systems based on genetic algorithms. In particular we study UCS, a rule-based classifier system which learns under a supervised learning scheme. We analyze UCS on an artificial domain with varying imbalance levels. We find UCS fairly sensitive to high levels of class imbalance, to the degree that UCS tends to evolve a simple model of the feature space classified according to the majority class. We analyze strategies for dealing with class imbalances, and find fitness adaptation based on class-sensitive accuracy a useful tool for alleviating the effects of class imbalances

[1]  Martin Houlind.  Translation and adaptation , 2001 .

[2]  Pier Luca Lanzi,et al.  An Analysis of Generalization in the XCS Classifier System , 1999, Evolutionary Computation.

[3]  Ester Bernadó-Mansilla,et al.  Accuracy-Based Learning Classifier Systems: Models, Analysis and Applications to Classification Tasks , 2003, Evolutionary Computation.

[4]  Dennis R. Durbin,et al.  A New Bootstrapping Method to Improve Classification Performance in Learning Classifier Systems , 2000, PPSN.

[5]  Xavier Llorà,et al.  XCS and GALE: A Comparative Study of Two Learning Classifier Systems on Data Mining , 2001, IWLCS.

[6]  Martin V. Butz,et al.  An algorithmic description of XCS , 2000, Soft Comput..

[7]  Tin Kam Ho,et al.  Domain of competence of XCS classifier system in complexity measurement space , 2005, IEEE Transactions on Evolutionary Computation.

[8]  Catherine Blake,et al.  UCI Repository of machine learning databases , 1998 .

[9]  H. Crichton-Miller.  Adaptation , 1926 .

[10]  Stewart W. Wilson.  Classifier Fitness Based on Accuracy , 1995, Evolutionary Computation.

[11]  Martin V. Butz,et al.  Data Mining in Learning Classifier Systems: Comparing XCS with GAssist , 2005, IWLCS.

[12]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[13]  Warren C. Eveland.  Progress in Theoretical Biology , 1968 .

[14]  M. Pelikán,et al.  Analyzing the evolutionary pressures in XCS , 2001 .

[15]  J. Holmes.  Differential Negative Reinforcement Improves Classifier System Learning Rate in Two-Class Problems with Unequal Base Rates , 1990 .

[16]  Nathalie Japkowicz,et al.  The class imbalance problem: A systematic study , 2002, Intell. Data Anal..

[17]  JapkowiczNathalie,et al.  The class imbalance problem: A systematic study , 2002 .

[18]  Martin V. Butz,et al.  Rule-Based Evolutionary Online Learning Systems: Learning Bounds, Classification, and Prediction , 2004 .

[19]  M. F.,et al.  Bibliography , 1985, Experimental Gerontology.

[20]  Tom Fawcett,et al.  Adaptive Fraud Detection , 1997, Data Mining and Knowledge Discovery.

[21]  Stewart W. Wilson.  Generalization in the XCS Classifier System , 1998 .

[22]  Tim Kovacs,et al.  What Makes a Problem Hard for XCS? , 2000, IWLCS.

[23]  David D. Lewis,et al.  A comparison of two learning algorithms for text categorization , 1994 .