13.07.2015 Views

The Generation of Fuzzy Rules from Decision Trees - Department of ...

The Generation of Fuzzy Rules from Decision Trees - Department of ...

The Generation of Fuzzy Rules from Decision Trees - Department of ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

5. SummaryA method for creating fuzzy rules <strong>from</strong> decision trees has been presented and analyzed. <strong>The</strong> methodtends to produce too many rules. <strong>The</strong> rules are generated directly <strong>from</strong> data. <strong>The</strong> control resulting <strong>from</strong>the rules suggests they can benefit <strong>from</strong> post generation tuning. An example using fuzzy entropy wasshown which can reduce the size <strong>of</strong> the generated decision tree and hence reduce the number <strong>of</strong> fuzzyrules created <strong>from</strong> the decision tree. This paper did not address how to create the actual rules <strong>from</strong>the decision tree that was obtained with the use <strong>of</strong> the fuzzy entropy measure. It is possible to createproportionally weighted classes based on the mixture <strong>of</strong> examples at the node or to create a new modifiedset <strong>of</strong> outputs sets based upon the final tree, for example. Both <strong>of</strong> these approaches are being pursued.Acknowledgements: This research was partially supported by the United States <strong>Department</strong> <strong>of</strong> Energythrough the Sandia National Laboratories LDRD program, contract number DE-AC04-76DO00789. Thanks toHamid Berenji who initially posed this problem and discussed an early attempt. Also to the NASA-Ames researchcenter who provided space during Hall’s sabbatical.References[1] A. Bensaid, L. Hall, J. Bezdek, L. Clarke, M. Silbiger, and et.al. Validity-guided (re)clustering for imagesegmentation. IEEE Transactions on <strong>Fuzzy</strong> Systems, 4(2):112–123, 1996.[2] H. Berenji. Machine learning in fuzzy control. In Iizuka 90, pages 231–234, 1990.[3] H. Berenji and P. Khedkar. Learning and tuning fuzzy logic controllers through reinforcements. IEEETransactions on Neural Networks, 3(5):724–740, 1992.[4] J. Bezdek and S. Pal, editors. <strong>Fuzzy</strong> Models for Pattern Recognition. IEEE Press, 1992.[5] G. Box and G. Jenkins. Time Series Analysis: Forecasting and Control. Holden Day, 1970.[6] R. Cannon, J. Dave, and J. Bezdek. Efficient implementation <strong>of</strong> the fuzzy c-means clustering algorithms.IEEE Transactions on Pattern Analysis and Machine Intelligence, 8:248–255, 1986.[7] Z. Chi and H. Yan. ID3-derived fuzzy rules and optimized defuzzification for handwritten numeral recognition.IEEE Transactions on <strong>Fuzzy</strong> Systems, 4(1):24–31, February 1996.[8] L. Hall and P. Lande. Generating fuzzy rules <strong>from</strong> data. In FUZZ-IEEE’96, pages 1757–1762, 1996.[9] L. Hall and P. Lande. Generating fuzzy rules <strong>from</strong> decision trees. In Seventh International <strong>Fuzzy</strong> SystemsAssociation World Conference 97, pages I–417–423, 1997.[10] J.-R. Jang. Anfis: Adaptive-network-based fuzzy inference systems. IEEE Transactions on Systems, Manand Cybernetics, 23(3):665–685, 1993.[11] J.-S. Jang and N. Gulley. <strong>Fuzzy</strong> Logic Toolbox for Use with Matlab. <strong>The</strong> MathWorks, Inc., 24 Prime ParkWay, Natick Mass., 1995.[12] L. Koczy and K. Hirota. Size reduction by interpolation in fuzzy rule bases. IEEE Transactions on Systems,Man and Cybernetics, 27(1):14–25, Feb. 1997.[13] P. Lande. Automated fuzzy controller learning and generation by rule extraction <strong>from</strong> classical decision trees.Master’s thesis, University <strong>of</strong> South Florida, Tampa, Fl. 33620, Dept. <strong>of</strong> CSE, 1995.[14] P. Murphy and M. Pazzani. Exploring the decision forest: An empirical investigation <strong>of</strong> occam’s razor indecision tree induction. Journal <strong>of</strong> Artificial Intelligence Research, 1:257–275, 1994.[15] S. Murthy, S. Kasif, and S. Salzberg. A system for induction <strong>of</strong> oblique decision trees. Journal <strong>of</strong> ArtificialIntelligence Research, 2:1–33, 1994.[16] O. Nelles, M. Fisher, and B. Muller. <strong>Fuzzy</strong> rule extraction by a genetic algorithm and constrained nonlinearoptimization. In Fifth IEEE International Conference on <strong>Fuzzy</strong> Systems, pages 213–219, 1996.[17] R. Osborne. <strong>Fuzzy</strong> clips 6.02a documentation. Technical Report http://ai.iit.nrc.ca/fuzzy/fuzzy.html, CanadianResearch Council, Institute for Information Technology, 1995.[18] I. Ozyurt and L. Hall. Uncertainty Analysis in Engineering, billal ayyub, ed. <strong>Fuzzy</strong> Genetic Algorithm BasedApproach to Machine Learning. Kluwer Academic, 1998.[19] J. Quinlan. Improved use <strong>of</strong> continuous attributes in c4.5. Journal <strong>of</strong> Artificial Intelligence Research, 4:77–90,1996.9

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!