A Comparison of Rule Matching Methods Used in AQ15 and LERS
This paper focuses on a performance comparison of two rule matching (classification) methods, used in data mining systems AQ15 and LERS. All rule sets used in our experiments were induced by the LERS (Learning from Examples using Rough Sets) system from ten typical input data sets. Then these rule sets were truncated using three different criteria: t-weight, u-weight and the strongest rule. The truncation process was performed using six different cut-off values for tweight, six different cut-off values for u-weight and using the strongest rule option. Hence for each of the input rule files thirteen truncated rule sets were created. Performance was measured by a classification error rate. The objective of this study was to determine the best overall method of classification and the best truncation option.
KeywordsKnowledge discovery and data mining rule induction classification systems rule truncation AQ15 LERS ten-fold cross validation Wilcoxon matched-pairs signed rank test.
Unable to display preview. Download preview PDF.
- 1.Bergadano, F., Matwin, S., Michalski, R. S., and Zhang, J.: Learning Two-Tiered Descriptions of Flexible Concepts: The POSEIDON System, Machine Learning 8 (1992), 5–43.Google Scholar
- 2.Booker, L. B., Goldberg, D. E., and Holland, J. F.: Classifier Systems and Genetic Algorithms. In Carbonell, J. G. (ed.): Machine Learning. Paradigms and Methods. The MIT Press (1990) 235–282.Google Scholar
- 4.Grzymala-Busse, J. W.: LERS—A System for Learning from Examples Based on Rough Sets. In: Slowinski, R. (ed.): Intelligent Decision Support. Handbook of Applications and Advances of the Rough Sets Theory. Kluwer Academic Publishers, Boston, MA (1992) 3–18.Google Scholar
- 5.Grzymala-Busse, J. W.: Managing Uncertainty in Machine Learning from Examples. Proc. of the Third Intelligent Information Systems Workshop, Wigry, Poland, June 6-11, 1994, 70–84.Google Scholar
- 6.Holland, J. H., Holyoak K. J., and Nisbett, R. E.: Induction. Processes of Inference, Learning, and Discovery. The MIT Press (1986).Google Scholar
- 7.Hamburg, M.: Statistical Analysis for Decision Making. Harcourt Brace Jovanovich Inc. (1983) Third Edition.Google Scholar
- 8.Kaufman, K. A. and Michalski, R. S.: The AQ18 System for Machine Learning: User’s Guide, Reports of Machine Learning and Inference Laboratory, MLI 00-3, George Mason University, Fairfax, VA, 2000.Google Scholar
- 9.Michalski, R. S.: A Theory and Methodology of Inductive Learning. In: Michalski, R. S., Carbonell, J. G., Mitchell T. M. (eds.): Machine Learning. An Artificial Intelligence Approach, Morgan Kauffman (1983) 83–134.Google Scholar
- 10.Michalski, R. S., Mozetic, I., Hong, J. and Lavrac, N.: The AQ15 Inductive Learning system: An Overview and Experiments. Department of Computer Science, University of Illinois, Rep. UIUCDCD-R-86-1260 (1986).Google Scholar
- 13.Reinke, R. E.: Knowledge Acquisition and Refinement Tools for the ADVISE META_EXPERT System, M. S. Thesis, Reports of the Intelligent Systems Group, ISG 84-4, UIUCDCS-F-84-921, Department of Computer Science, University of Illinois, Urbana, July 1984.Google Scholar
- 14.Slowinski, R. and Stefanowski, J.: ‘RoughDAS’ and ‘RoughClass’ Software Implementations of the Rough Set Approach. In: Slowinski, R. (ed.): Intelligent Decision Support. Handbook of Applications and Advances of the Rough Sets Theory. Kluwer Academic Publishers, Boston, MA (1992) 445–456.Google Scholar
- 15.Weiss, S. M. & Kulikowski, C. A.: Computer Systems That Learn: Classification and Prediction Methods from Statistics, Neural Nets, Machine Learning, and Expert Systems. Morgan Kaufmann Publishers (1991).Google Scholar