The Extraction of Elliptical Rules from the Trained Radial Basis Function Neural Network

Andrey Bondarenko, Arkady Borisov

Abstract


The paper describes an algorithm for approximation of trained radial basis function neural network (RBFNN) classification boundary with the help of elliptic rules. These rules can later be translated into IF-THEN form if required. We provide experimental results of the algorithm for a two-dimensional case. Currently, neural networks are not widely used and spread due to difficulties with the interpretation of classification decision being made. The formalized representation of decision process is required in many mission critical areas, such as medicine, nuclear energy, finance and others.

Keywords:
Knowledge acquisition; optimization; radial basis function networks

Full Text:

 PDF

References


J. Moody and C.J.Darken, Fast Learning in networks of locally tuned processing units, Neural Computation, 1, pp.281-294, 1989.

Poggio and F.Girosi, Networks for approximation and learning, Proc. IEEE 78(9), pp.1484-1487, 1990.

S. Chen, X. Hong, B.L. Luk and C.J. Harris, Construction of tunable radial basis function networks using orthogonal forward selection, IEEE Trans. Systems, Man, and Cybernetics, Part B, Vol.39, No.2, pp.457- 466, 2009.

A. Bondarenko, V. Jumutc, Extraction of Interpretable Rules from Piecewise-Linear Approximation of a Nonlinear Classifier Using Clustering-Based Decomposition, Cambridge, AIKED 2011.

H.Nunez, C. Angulo and A.Catala, Rule extraction from support vector machines. Proceedings of the 10th European Symposium on Artificial Neural Networks, pp.335-343, Bruges, Belgium, April24-26, 2002.

S. Boyd, L. Vandenberghe, Convex Optimization. Cambridge University Press, 2004.

A.Frank and A.Asuncion, UCI Machine Learning Repository, [http://archive.ics.uci.edu/ml], School of Information and Computer Science, Irvine, University of California. [Accessed Sept. 29, 2012].


Refbacks

  • There are currently no refbacks.


Copyright (c) 2012 Andrey Bondarenko, Arkady Borisov

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.