Abstract
Recently, neural networks have been applied to many medical diagnostic problems because of their appealing properties, robustness, capability of generalization and fault tolerance. Although the predictive accuracy of neural networks may be higher than that of traditional methods (e.g., statistical methods) or human experts, the lack of explanation from a trained neural network leads to the difficulty that users would hesitate to take the advise of a black box on faith alone. This paper presents a class of composite neural networks which are trained in such a way that the values of the network parameters can be utilized to generate If-Then rules on the basis of preselected meaningful coordinates. The concepts and methods presented in the paper are illustrated through one practical example from medical diagnosis.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
L. M. Fu, “Rule learning by searching on adapted nets,” AAAI'91, pp. 590–595, 1991.
L. Bochereau and P. Bourgine, “Extraction of sermantic features and logical rules from a multilayer neural network,” Proceedings of IJCNN 90, Washington D.C., pp. 579–582, 1991.
R. M. Goodman, C. M. Higgins, and J. W. Miller, “Rule-based neural networks for classification and probability estimation,” Neural Computation, Vol. 4, pp. 781–804, 1992.
S. I. Gallant, “Connectionist expert systems,” Communication of the ACM, Vol. 31, pp. 152–169, 1988.
S. I. Gallant, Neural Network Learning and Expert Systems. Cambridge, Massachusetts: MIT Press, 1993.
L. A. Zadeh, “Fuzzy sets,” Information and Control, Vol. 8, pp. 338–353, 1965.
P. K. Simpson, “Fuzzy min-max neural networks-part 1: Classification,” IEEE Trans. on Neural Networks, Vol. 3, pp. 776–786, 1965.
S. Abe and M. S. Lan, “A method for fuzzy rules extraction directly from numerical data and its application to pattern recognition,” IEEE Trans. on Fuzzy Systems, Vol. 3, No. 1, pp. 18–28, 1995.
H. Takagi and I. Hayashi, “NN-driven fuzzy reasoning,” Int. J. Approximate Reasoning, Vol. 5, No. 3, pp. 191–212, 1991.
N. DeClaris and M.-C. Su, “A novel class of neural networks with quadrastic junctions,” IEEE Int. Conf. on Systems, Man, and Cybernetics, pp. 1557–1562, 1965. (Received the 1992 Franklin V. Taylor Award)
N. DeClaris and M.-C. Su, “Introduction to the theory and application of neural networks with quadratic junctions,” IEEE Int. Conf. on System, Man, and Cybernetics, pp. 1320–1325, 1992.
N. DeClaris and M.-C. Su, “Learning and generalization in neural networks with quadratic junctions,” symposium on Implicit and Nonlinear Systems, pp. 57–65, 1992.
T. Poggio and F. Girosi, “Networks for approximation and learning,” Proc. of the IEEE, Vol. 78, No. 9, pp. 1481–1497, 1990.
M.-C. Su, C. J. Kao, K. M. Liu, and C. Y. Liu, “Rule extracting using a novel class of fuzzy degraded hyperellipsoidal composite neural networks,” Fuzz-IEEE/IFES'95, pp. 233–238, 1990.
B. C. Hansen and N. L. Bodkin, “Heterogeneity of insulin responses: phase leading to type 2 (noninsulin) diabetes mellitus in the Rhesus monkey,” Diabetologia, Vol. 29, pp. 713–719, 1990.
M.-C. Su, “Use of neural networks as medical diagnosis expert systems,” Comput. Biol. Med., Vol. 24, No. 6, pp. 419–429, 1990. Also appeared in 1996 IMIA Yearbook of Medical Informatics, pp. 439–449, Schattauer, 1996.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Su, MC., Chang, HT. Extracting Rules from Composite Neural Networks for Medical Diagnostic Problems. Neural Processing Letters 8, 253–263 (1998). https://doi.org/10.1023/A:1009681803460
Issue Date:
DOI: https://doi.org/10.1023/A:1009681803460