Rule extraction from approximating neural network - a real challenge for evolutionary method

• Autorzy: Markowska-Kaczmar U. , Mularczyk K.
• Języki publikacji: EN

Abstrakty

EN

The paper describes a modification of the MulGex method in order to use it to extract rules from approximating neural network. Originally the method was designed to extract prepositional rules from classification neural network. The rules are searched by evolutionary algorithms working on two levels. The rules are optimized using the Pareto approach. The main principle referring to premise part of a rule has been unchanged but the form of conclusion instead of the class label describes a formula which can be a linear function and is encoded as a list of coefficients or it takes a form of a tree whose inner nodes contain functions and operators, and leaves - identifiers of attributes and numeric constants. Although the results obtained in the experiments for three different data sets can be assumed as satisfactory, some changes improving MulGex efficiency are proposed at the end.

Słowa kluczowe

EN

rule extraction; neural network; evolutionary algorithm

PL

sieć neuronowa; algorytm ewolucyjny; aproksymacja

• Wydawca: Oficyna Wydawnicza Politechniki Wrocławskiej
• Czasopismo: Systems Science
• Rocznik: 2006
• Tom: Vol. 32, no 4
• Strony: 35--43
• Opis fizyczny: Bibliogr. 17 poz., rys.

Twórcy

autor

Markowska-Kaczmar U.

autor

Mularczyk K.

• Institute of Applied Informatic, Wroclaw University of Technology

Bibliografia

• [1] Andrews R., Diedrich J., Tickle A., Survey and critique of techniques for extracting rules from trained artificial neural networks, Knowledge-Based Systems. Vol. 8, Issue 6. December, 1995, pp. 373-389.
• [2] Craven M., Shavlik J, Using sampling and queries to Extract Rules from Trained Neural Networks, Proc. 11th Int. Conf. Machine Learning, San Mateo, Morgan Kaufmann, 1994, pp. 3-45.
• [3] Coello C., Van Veldhuizen D., Lament G.B., Evolutionary Algorithms for Solving Multi-Objective Problems, Kluwer Academic Publisher, New York 2002.
• [4] Darbari A., Rule Extraction from Trained ANN: A Survey, Tech. Rep. WV-2000-03, Knowledge Representation and Reasoning Group, Department of Computer Science, Dresden University of Technology, Dresden, Germany, 2000.
• [5] Fonseca C, Fleming P., Genetic algorithms for multiobjective optimization: Formulation, discussion and generalization, Proc. Fifth Int. Conf. on Genetic Algorithms, San Mateo, California, 1993, pp. 416-423.
• [6] Ishibuchi H., Murata T., Türksen I.B., Single-objective and two-objective genetic algorithms for selecting linguistic rules for pattern classification problems, Fuzzy Sets and Systems 89, 1997, pp. 135-150.
• [7] Jacobsson H., Rule extraction from recurrent neural networks: A taxonomy and review. Neural Computation, 17(6), 2005. pp. 1223-1263.
• [8] Jin Y., Sendhoff B., Extracting Interpretable Fuzzy Rules from RBF Networks Neural, Processing Letters, Vol. 17, Issue 2, 2003, pp. 149-164.
• [9] Koza J., Genetic Programming, MIT Press, London 1996.
• [10] Malone J., McGarry K., Wermter S., Bowerman Ch., Data mining using rule extraction from Kohonen self-organising maps, Neural computing and applications, Vol. 15, No. 1, 2006, pp. 9-17.
• [11] Markowska-Kaczmar U., Mularczyk K., Gabased Pareto optimization for rule extraction from neural networks, Multiobjective Machine Learning. Studies in Computational Intelligence, 2006.
• [12] Markowska-Kaczmar U., Trelak W., Fuzzy logic and evolutionary algorithm - two techniques in rule extraction from neural networks, Neurocomputing, 63, 2005, pp. 359-379.
• [13] McDonald G.C., Schwing R., Instabilities of regression estimates relating air pollution to mortality, Technometrics, 15, 1973, pp. 463-482.
• [14] Saito K., Nakano R., Law Discovery using Neural Network, Proc. IEEE/IAFE/INFORM Conf. Computational Intelligence for Financial Engineering, 2000, pp. 209-211.
• [15] Setiono R., Leow W., Zurada J.M., Extraction of rules from artificial neural networks for nonlinear regression, IEEE Trans. Neural Networks, Vol. 13, No. 3, 2002, pp. 564-577.
• [16] Wallace M., Tsapatsoulis N., Combining GAs and RBF neural networks for fuzzy rule extraction from numerical data. Int. Conf. Artificial Neural Networks. Lecture notes comput. sci., No. 15, Warsaw, Poland 2005, Vol. 3697, pp. 521-526.
• [17] Weijters A., Paredis J., Rule induction with genetic sequential covering algorithm (geseco), Proc. 2nd ICSC Symp. Engineering of Intelligent Systems, (EIS 2000), 2000, pp. 245-251.