Abstract

Inspired originally by the Learnable Evolution Model(LEM) a new presents of new classification algorithm called (LEM+ID3), which is based on the techniques from the learnable evolution models (LEM) to enhance convergence and accuracy of the algorithm and use of ID3 in order to construct the tree used in classification. In this paper a new version of LEM which convert LEM from optimization domain to classification domain and then examine the feature extraction problems and show that learning evolutional can significantly enhance the performance of pattern recognition systems with simple classifiers. This model is applied to real world datasets from the UCI Machine Learning databases to verify proposed approach and compare it with other convention classifiers. The conclusion is this algorithm is able to produce classifiers of superior (or equivalent) performance to the conventional classifiers examined Also time taken to reach near optimum accuracy.

References

• R. S. Michalski, "Learnable Evolutionary Model: Evolutionary Processes Guided by Machine Learning". Machine Learning, Vol. 38, pp. 9-40, 2000.
• Wnek, J. , Kaufmann, K. Bloedorn, E. , Michalski R. S. Inductive Learning System AQ15c: the method and users guide. Reports of the Machine Learning and Inference Laboratory, MLI95-4, George Mason University, Fairfax, VA, USA,1995.
• Jourdan,L. , Corne, D. , Savic, D. , Walters, G. Hybridising rule induction and multiobjective evolutionary search for optimizing water distribution systems, in Proc of the 4th Hybrid Intelligent Systems conference, published in 2005 by IEEE Com- puter Society Press. pp. 434-439, ISBN 0-7695-1916-4. , (2005)
• Larranaga, P. , Lozano, J. A. (eds) (2002) Stimulation of Distribution Algorithms: A New Tool for Evolutionary Computation, Kluwer Academic Publishers.
• J. Wojtusiak and R. S. Michalski,"The LEM3 Implementation of Learnable Evolution Model and Its Testing on Complex Function Optimization Problems", Proceedings of Genetic and Evolutionary Computation Conference, GECCO 2006, Seattle, WA, July 8-12, 2006.
• M. Ebner and A. Zell, Evolving a task specific image operator, in Joint Proceedings of the 1st EuropeanWorkshop on Evolutionary Image Analysis,Signal Processing and Telecommunications (EvoIASP'99 and EuroEcTel'99), Goteborg, Sweden (1999) pp. 74–89
• D. Michie, D. J. Spiegelhalter and C. C. Taylor, Machine Learning, Neural and Statistical Classification, Ellis Horwood, Upper Saddle River,NJ, 1994
• T. Lim, W. Loh and Y. Shih, A comparison of prediction accuracy, complexity, and training time of thirty-three old and new classification algorithms, Machine Learning, 40 (2000) 203–228
• S. Parsons, "Introduction to Machine Learning by Ethem Alpaydin," The Knowledge Engineering Review, vol. 20, no. 4, pp. 432-433, 2005.
• J. R. Quinlan, "Induction of Decision Trees " Machine Learning, vol. 1, no. 1, pp. 81-106, 1986.
• O. L. Mangasarian, W. N. Street, and W. H. Wolberg, Breast cancer diagnosis and prognosis via linear programming, Operations Research 43 (1995) 570–577
• J. Wojtusiak and R. S. Michalski,"The LEM3 Implementation of Learnable Evolution Model and Its Testing on Complex Function Optimization Problems", Proceedings of Genetic and Evolutionary Computation Conference, GECCO 2006, Seattle, WA, July 8-12, 2006.
• L. Jourdan, D. Corne, D. Savic, G. Walters (2005) Hybridising rule induction and multiobjective evolutionary search for optimizing water distribution systems, in Proc of the 4th Hybrid Intelligent Systems conference, published in 2005 by IEEE Computer Society Press. Pp. 434-439, ISBN 0-7695-1916-4.
• I. H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools, 2nd ed. , Morgan Kaufmann, San Francisco, CA, 2005
• T. Lim, W. Loh and Y. Shih, A comparison of prediction accuracy, complexity, and training time of thirty-three old and new classification algorithms, Machine Learning, 40 (2000) 203–228.