Abstract   This  paper  reports  on  the  development  of  a  new  hybr id  architecture  that  integrates  Learning  Classifier  Systems  (LCS)  with  Rete-based  production  systems  inference  engine  to  improve  the  performance  of  the  process  of  compacting  LCS  generated rules. While LCS is responsible for gener ating a complete ruleset from a given breast cancer pathological data-set,  an adapted Rete-based inference engine has been int egrated for the efficient extraction of a minimal and representative ruleset  from  the  original  generated  ruleset.  This  has  resul ted  in  an  architecture  that  is  hybrid,  efficient,  component-based,  elegant,  and  extensible.  Also,  this  has  demonstrated  signifi cant  savings  in  computing  the  match  phase  when  buil ding  on  the  two  main  features of the Rete match algorithm, namely struct ural similarity and temporal redundancy. Finally, this architecture  may be  considered as a new platform for research on compac tion of LCS rules using Rete-based inference engines.  

References

[1] Bernado E., Llor X., and Garrell J., XCS and GALE: A Comparative Study of Two Classifier Systems with Six other Learning Algorithms on Classification Tasks, in Proceedings of the International Workshop on Learning Classifier System , London, pp. 115-132, 2001.

[2] Bernado E., Llora X., and Garrell J., XCS and GALE: A Comparative Study of Two Learning Classifier Systems on Data Mining, in Proceedings of the 4 th International Workshop, Advances in Learning Classifier Systems , Berlin, Germany, vol. 2321 pp. 115-132, 2002.

[3] Blake C. and Merz C., UCI Repository of Machine Learning Databases, available at: http://www.ics.uci.edu/~mlearn/MLRepository. html, last visited 1998.

[4] Bonelli P., Parodi A., Sen S., and Wilson S., NEWBOOLE: A Fast GBML System, in Proceedings of International Conference on Machine Learning , USA, pp. 153-159, 1990.

[5] Bull L., Applications of Learning Classifier Systems , Springer, 2004.

[6] Dixon P., Corne D., and Oates M., A Ruleset Reduction Algorithm for the XCS Learning Classifier System, in Proceedings of the 5 th International Workshop, Learning Classifiers Systems , Spain, pp. 20-29, 2003.

[7] Forgy C., Rete: A Fast Algorithm for the Many Pattern/Many Object Pattern Match Problem, Artificial Intelligence , vol. 19, no. 1, pp. 17-37, 1982.

[8] Friedman-Hill E., Jess in Action Java Rule-Based Systems , USA, 2008.

[9] Fu C. and Davis L., A Modified Classifier System Compaction Algorithm, in Proceedings of the Genetic and Evolutionary Computation Conference , USA, pp. 920-925, 2002.

[10] Giarratano J. and Riley G., Expert Systems: Principles and Programming , Course Technology, USA, 1998.

[11] Gonzalez A. and Dankel D., The Engineering of Knowledge-Based Systems: Theory and Practice , Prentice-Hall, USA, 1993.

[12] Holland J. and Reitman J., Cognitive Systems Based on Adaptive Algorithms, in Proceedings of ACM SIGART Bulletin, USA, pp. 49-49, 1977.

[13] Holland J., Progress in Theoretical Biology IV , Academic Press, pp. 263-93, 1976.

[14] Holland J., Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , Bradford Book, 1992.

[15] Holmes J., Lanzi P., Stolzmann W., and Wilson S., Learning Classifier Systems: New Models, Successful Applications, Information Processing Letters , vol. 82, no. 1, pp. 23-30, 2002.

[16] Kovacs T., XCS Classifier System Reliably Evolves Accurate, Complete, and Minimal Representations for Boolean Functions, in Proceedings of Soft Computing in Engineering Design and Manufacturing , England, UK, pp. 59- 68, 1997.

[17] Mangasarian O. and Wolberg H., Cancer Diagnosis via Linear-Programming, SIAM News , vol. 23, no. 5, pp. 1-18, 1990. A New Hybrid Architecture for the Discovery and Compaction of Knowledge: Breast Cancer 123

[18] Odeh M., Concurrent Object-Oriented Execution of OPS5 Production Systems , University of Bath, UK, 1993.

[19] Saxon S. and Barry A., XCS and the Monk's Problems, in Proceedings of Learning Classifier Systems , Berlin, Germany, vol. 1813, pp. 223- 242, 2000.

[20] Shortliffe E., Computer-Based Medical Consultations: MYCIN , Elsevier, USA, 1976.

[21] Soloway E., Bachant J., and Jensen K., Assessing the Maintainability of XCON-IN- RIME: Coping with the Problems of a VERY Large Rule-Base, in Proceedings of the International Conference on Artificial Intelligence , USA, pp. 825-829, 1987.

[22] Wilson S., Compact Rulesets from XCSI, in Proceedings of the 4 th International Workshop on Learning Classifier Systems , London, UK, pp. 197-210, 2001.

[23] Wilson S., Knowledge Growth in an Artificial Animal, in Proceedings of the 4 th Yale Workshop on Applications of Adaptive Systems Theory , USA, pp. 98-104, 1985.

[24] Wilson S., Quasi-Darwinian Learning in a Classifier System, in Proceedings of the 4 th International Workshop on Machine Learning , USA, pp. 59-65, 1987.

[25] Wilson S., Mining Oblique Data with XCS, in Proceedings of the 3 rd International Workshop, Learning Classifier Systems , France, vol. 1996, pp. 158-174, 2001.

[26] Wilson S., ZCS: A Zeroth Level Classifier System, Evolutionary Computation , vol. 2, no. 1, pp. 1-18, 1994.

[27] Wilson S., Classifier Fitness Based on Accuracy, Evolutionary Computation , vol. 3, no. 2, pp. 149-175, 1995.

[28] Wilson S. and Goldberg D., A Critical Review of Classifier Systems, in Proceedings of the 3 rd International Conference on Genetic Algorithms , USA, pp. 244-255, 1989.

[29] Wilson S., Get Real! XCS with Continuous- Valued Inputs, in Proceedings of Learning Classifier Systems , Berlin, Germany, vol. 1813, pp. 209-219, 2000.

[30] Wyatt D., Bull L., and Parmee I., Building Compact Rulesets for Describing Continuous- Valued Problem Spaces using a Learning Classifier System, in Proceedings of Adaptive Computing in Design and Manufacture VI , London, UK, pp. 235-248, 2004. Faten Kharbat is an assistant professor in Artificial Intelligence at the Al-Ain University, Abu Dhabi Campus, UAE. She holds PhD degree in computer science from the University of the West of England, UK, in 2006. Her main research interest are learnin g classifier systems, applying ontology into translat ion engines, knowledge based systems, applying data mining techniques to marketing, and recently is involved in quality of higher education. Mohammed Odeh is senior lecturer in software engineering and associate of the Complex Cooperative Systems Centre at the University of West of England, Bristol, UK. He holds PhD degree in computer science from the University of Bath, 1993 in addition to PG Cert in Higher Education and membership of ACM and ILT. He has more than 20 years of experience including extensive project management experience in planning and leading a range of IT related projects in addit ion to management posts. He is the UWE principal investigator on the Onto REM knowledge exchange partnership with Airbus, and the SoAgile project be ing reviewed. His main research interests are bridging the gap between business process models and system models, ontology-driven requirements engineering, semantic and service-oriented software engineering, software cost estimation, grid computing, and knowledge management. His applied software engineering experience has been associated with banking, aerospace manufacturing, and medical informatics. Larry Bull is a professor of artificial intelligence and based in the Department of Computer Science & Creative Technologies at UWE. His research interests are in intelligent and unconventional systems, with an emphasis on evolution. He is the founding Editor-in-Chief of th e Springer journal Evolutionary Intelligence.