Towards clustering with XCS

Tamee, Kreangsak; Bull, Larry; Pinngern, Ouen

doi:10.1145/1276958.1277326

Cited by 24 publications

(9 citation statements)

References 20 publications

(21 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Hence the set pressure encourages the evolution of rules which cover many data points and the fitness pressure acts as a limit upon the separation of such data points, i.e., the error. Tammee et al [2006] began by using a slightly simplified version of XCS as the underlying LCS (YCS) [Bull, 2005], but found that XCS's relative accuracy fitness function was more effective than a function directly inversely proportional to error [Tammee et al, 2007]  is set high, e.g., 0.1, in less-separated data the contiguous clusters are covered by the same rules. They therefore developed an adaptive threshold parameter scheme which uses the average error of the current [M]:…”

Section: Xcsc: Unsupervised Learningmentioning

confidence: 99%

A brief history of learning classifier systems: from CS-1 to XCS and its variants

Bull

2015

Evol. Intel.

Self Cite

View full text Add to dashboard Cite

The legacy of Wilson's XCS is that modern Learning Classifier Systems can be characterized by their use of rule accuracy as the utility metric for the search algorithm(s) discovering useful rules. Such searching typically takes place within the restricted space of co-active rules for efficiency. This paper gives an historical overview of the evolution of such systems up to XCS, and then some of the subsequent developments of Wilson's algorithm to different types of learning.

show abstract

Section: Xcsc: Unsupervised Learningmentioning

confidence: 99%

A brief history of learning classifier systems: from CS-1 to XCS and its variants

Bull

2015

Evol. Intel.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Generally, they can be characterised by handling sequential decision tasks with a rule-based representation and by the use of evolutionary computation methods (for example, [167,95]), although some variants also perform supervised learning (for example, [161]) or unsupervised learning (for example, [211]), or do not rely on evolutionary computation (for example, [89]). Generally, they can be characterised by handling sequential decision tasks with a rule-based representation and by the use of evolutionary computation methods (for example, [167,95]), although some variants also perform supervised learning (for example, [161]) or unsupervised learning (for example, [211]), or do not rely on evolutionary computation (for example, [89]).…”

Section: Learning Classifier Systemsmentioning

confidence: 99%

Design and Analysis of Learning Classifier Systems

2008

Studies in Computational Intelligence

View full text Add to dashboard Cite

VI Foreword programming or reinforcement learning. Its close relation to XCS in this book results from the authors commitment to an LCS model structure that relates to XCS, but nothing speaks against applying the same approach to greatly different model types, resulting in different, potentially novel, LCS.While its connection to Pittsburgh-style LCS is straightforward and clearly established in this work, using the same approach for the design of Michiganstyle LCS remains a significant future challenge. Also, it will be interesting to see how the author's theoretical basis for reinforcement learning is built upon in future LCS, and how LCS designed by the model-based approach perform in comparison to currently existing LCS.Overall, the work is elegant and approaches LCS from a refreshingly different perspective. It's also stylistically pretty novel for LCS work -but that's surely healthy!

show abstract

“…CRA2 yielded similar performance to CRA but ran much faster when tested on the same dataset. Other similarly themed LCS rule compaction strategies include an approach for continuous-valued problem spaces (Wyatt et al, 2004), an approach for online rule compaction (Gao et al, 2006), an approach for clustering in XCS (Tamee et al, 2007), an approach which adds entropy calculation (Kharbat et al, 2008), and an approach designed for fuzzy rule representations (Shoeleh et al, 2011).…”

Section: Introductionmentioning

confidence: 99%

Rapid Rule Compaction Strategies for Global Knowledge Discovery in a Supervised Learning Classifier System

Tan¹,

Moore²,

Urbanowicz³

2013

Advances in Artificial Life, ECAL 2013

View full text Add to dashboard Cite

Michigan-style learning classifier systems have availed themselves as a promising modeling and data mining strategy for bioinformaticists seeking to connect predictive variables with disease phenotypes. The resulting 'model' learned by these algorithms is comprised of an entire population of rules, some of which will inevitably be redundant or poor predictors. Rule compaction is a post-processing strategy for consolidating this rule population with the goal of improving interpretation and knowledge discovery. However, existing rule compaction strategies tend to reduce overall rule population performance along with population size, especially in the context of noisy problem domains such as bioinformatics. In the present study we introduce and evaluate two new rule compaction strategies (QRC, PDRC) and a simple rule filtering method (QRF), and compare them to three existing methodologies. These new strategies are tuned to fit with a global approach to knowledge discovery in which less emphasis is placed on minimizing rule population size (to facilitate manual rule inspection) and more is placed on preserving performance. This work identified the strengths and weaknesses of each approach, suggesting PDRC to be the most balanced approach trading a minimal loss in testing accuracy for significant gains or consistency in all other performance statistics.

show abstract

Towards clustering with XCS

Cited by 24 publications

References 20 publications

A brief history of learning classifier systems: from CS-1 to XCS and its variants

A brief history of learning classifier systems: from CS-1 to XCS and its variants

Design and Analysis of Learning Classifier Systems

Rapid Rule Compaction Strategies for Global Knowledge Discovery in a Supervised Learning Classifier System

Contact Info

Product

Resources

About