Learning Classifier Systems
The general idea of a classifier system and its tripartite structure have already been introduced in §1.5. We begin this Chapter by briefly outlining various types of classifier systems and reviewing the representation of rules in them. Following this we proceed to describe two classifier systems, accuracy-based XCS and its strength-based twin SB-XCS, in detail sufficient to allow implementation. We make a brief empirical comparison of them in §2.5, and will conduct more extensive comparisons and experiments in later Chapters. Through- out the thesis these two systems will be used as references in our reasoning about strength and accuracy-based systems.
KeywordsClassifier System Action Selection System Prediction Sequential Task Fitness Calculation
Unable to display preview. Download preview PDF.