Classifier systems: a useful approach to machine learning?

Classi er systems are sub-symbolic or dynamic approaches to machine learning. These systems have been studied rather extensively. In this thesis some theoretical results about the long-term behaviour and the computational abilities of classi er systems are derived. Then some experiments are undertaken. The rst experiment entails the implementation of a simple logic function, a multiplexer in a simple classi er system. It is shown that this task can be learned very well. The second task that is taught to the system is a mushroom-classi cation problem that has been researched with other learning systems. It is shown that this task can be learned. The last problem is the parity problem. First it is shown that this problem does not scale linearly with its number of bits in a straightforward classi er system. An attempt is made to solve it with a multilayer classi er-system, but this is found to be almost impossible. Explanations are given of why this should be the case. Then some thought is given to analogies between classi er systems and neural networks. It is indicated that there are mappings between certain classi er systems and certain neural networks. It is suggested that this is a main concern for future classi er systems research.

[1]  Dean Pomerleau,et al.  ALVINN, an autonomous land vehicle in a neural network , 2015 .

[2]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[3]  John H. Miller,et al.  Emergent behavior in classifier systems , 1990 .

[4]  Herbert A. Simon,et al.  WHY SHOULD MACHINES LEARN , 1983 .

[5]  J. Davies,et al.  Molecular Biology of the Cell , 1983, Bristol Medico-Chirurgical Journal.

[6]  John R. Anderson,et al.  MACHINE LEARNING An Artificial Intelligence Approach , 2009 .

[7]  John J. Grefenstette,et al.  Genetic algorithms and their applications , 1987 .

[8]  Dean A. Pomerleau,et al.  What's hidden in the hidden layers? , 1989 .

[9]  H. L. Cather Techniques in Computational Learning—an Introduction , 1993 .

[10]  Marco Dorigo,et al.  Message-Based Bucket Brigade: An Algorithm for the Apportionment of Credit Problem , 1991, EWSL.

[11]  Daniele Montanari,et al.  Learning and bucket brigade dynamics in classifier systems , 1990 .

[12]  Emil L. Post Formal Reductions of the General Combinatorial Decision Problem , 1943 .

[13]  K. Dejong,et al.  An analysis of the behavior of a class of genetic adaptive systems , 1975 .

[14]  Zbigniew Michalewicz,et al.  Genetic Algorithms + Data Structures = Evolution Programs , 1992, Artificial Intelligence.

[15]  Terrence J. Sejnowski,et al.  Parallel Networks that Learn to Pronounce English Text , 1987, Complex Syst..

[16]  Lashon B. Booker,et al.  Improving the Performance of Genetic Algorithms in Classifier Systems , 1985, ICGA.

[17]  Pat Langley,et al.  Trading Off Simplicity and Coverage in Incremental concept Learning , 1988, ML.

[18]  J. C. Schlimmer,et al.  Concept acquisition through representational adjustment , 1987 .

[19]  Robert E. Smith,et al.  Is a Learning Classifier System a Type of Neural Network? , 1994, Evolutionary Computation.

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

[21]  William Colgrove Intelligent user interfaces and the Internet , 1995 .

[22]  Patrick K. Simpson,et al.  Artificial Neural Systems: Foundations, Paradigms, Applications, and Implementations , 1990 .

[23]  W. Leech A rule-based process control method with feedback. , 1987, ISA transactions.

[24]  Hayong Zhou Classifier Systems with Long Term Memory , 1985, ICGA.

[25]  Anders Krogh,et al.  Introduction to the theory of neural computation , 1994, The advanced book program.

[26]  Thomas H. Westerdale,et al.  The Bucket Brigade Is Not Genetic , 1985, International Conference on Genetic Algorithms.

[27]  N. Pierce Origin of Species , 1914, Nature.