Abstract
The concept of combining modular neural networks has been recently exploited as a new direction for the development of highly reliable neural network systems in the area of pattern classification. In this paper we present an efficient method for combining the modular networks based on fuzzy logic, especially the fuzzy integral. This method nonlinearly combines objective evidences, in the form of network outputs, with subjective evaluation of the reliability of the individual neural networks. Also, for more effective aggregation, we adopt the extension of the fuzzy integral with ordered weighted averaging operators. The experimental results with the recognition problem of on-line handwriting characters show that the performance of individual networks could be improved significantly.
This work was partly conducted while at: KAIST Computer Science Department, 373-1 Koosung-dong, Yoosung-ku, Taejeon 305–701, Korea, sbchogorai.kaist.ac.kr; and ATR Human Information Processing Research Laboratories, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan, sbchohip.atr.co.jp
Preview
Unable to display preview. Download preview PDF.
References
Hansen, L.K., Salamon, P.: Neural network ensembles. IEEE Trans. Patt. Anal. Mach. Inte. 12 (1990) 993–1001
Scofield, C., Kenton, L., Chang, J.: Multiple neural net architectures for character recognition. Proc. Compcon. (1991) 487–491 (San Francisco, CA, IEEE Computer Society Press)
Jacobs, R.A., Jordan, M.I., Nowlan, S.J., Hinton, G.E.: Adaptive mixtures of local experts. Neural Comp. 3 (1991) 79–87
Wolpert, D.: Stacked generalization. Neural Net. 5 (1992) 241–259
Cho, S.-B., Kim, J.H.: Two design strategies of neural network for complex classification problems. Proc. 2nd Int. Conf. Fuzzy Logic & Neural Net. (1992) 759–762
Perrone, M.P., Cooper, L.N.: When networks disagree: Ensemble methods for hybrid neural networks. Neural Net. for Speech and Image Proc. (1993) (Mammone, R.J. ed., Chapman-Hill, London)
Hampshire, J.B. II, Waibel, A.: Connectionist architectures for multi-speaker phoneme recognition. Adv. Neural Info. Proc. Syst 2 (1990) 203–210
Xu, L., Krzyzak, A., Suen, C.Y.: Methods of combining multiple classifiers and their applications to handwriting recognition. IEEE Trans. Syst. Man. Cyber. 22 (1992) 688–704
Benediktsson, J.A., Swain, P.H.: Consensus theoretic classification methods. IEEE Trans. Syst. Man. Cyber. 22 (1992) 418–435
Cho, S.-B., Kim, J.H.: Combining multiple neural networks by fuzzy integral for robust classification. IEEE Trans. Syst. Man. Cyber. 25 (1995) 380–384
Richard, M.D., Lippmann, R.P.: Neural network classifiers estimate Bayesian a posteriori probabilities. Neural Comp. 3 (1991) 461–483
Shlien, S.: Multiple binary decision tree classifiers. Patt. Recog. 23 (1990) 757–763
Sugeno, M.: Fuzzy measures and fuzzy integrals: A survey. Fuzzy Automata Dec. Proc. (1977) 89–102 (Amsterdam: North Holland)
Leszeynski, K., Penczek, P., Grochulskki, W.: Sugeno's fuzzy measures and fuzzy clustering. Fuzzy Sets Syst. 15 (1985) 147–158
Tahani, H., Keller, J.M.: Information fusion in computer vision using the fuzzy integral. IEEE Trans. Syst. Man. Cyber. 20 (1990) 733–741
Yager, R.R.: Element selection from a fuzzy subset using the fuzzy integral. IEEE Trans. Syst. Man. Cyber. 23 (1993) 467–477
Yager, R.R.: On ordered weighted averaging aggregation operators in multicriteria decisionmaking. IEEE Trans. Syst. Man. Cyber. 18 (1988) 183–190
Tappert, C.C., Suen, C.Y., Wakahara, T.: The state of the art in on-line handwriting recognition. IEEE Trans. Patt. Anal. Mach. Intel. 12 (1990) 787–808
Freeman, H.: Computer processing in line drawing images. Comp. Surv. 6 (1974) 57–98
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Cho, SB. (1995). A neuro-fuzzy architecture for high performance classification. In: Furuhashi, T. (eds) Advances in Fuzzy Logic, Neural Networks and Genetic Algorithms. WWW 1994. Lecture Notes in Computer Science, vol 1011. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-60607-6_6
Download citation
DOI: https://doi.org/10.1007/3-540-60607-6_6
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-60607-9
Online ISBN: 978-3-540-48457-8
eBook Packages: Springer Book Archive