Abstract
Circular Cascade Correlation (CCC) is an improvement of the well known Cascade Correlation algorithm: the modification consists in adding a new input, whose value is the mean of the squared original inputs. This addition allows hidden units to detect closed subspaces of the input space besides the usual open ones and it increases the representation capability of the network with a moderate growth of its complexity. A convergence theorem is proved, which shows that the number of hidden units needed by CCC to learn a finite training set is equal or lower than its cardinality.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Drago G.P. and Ridella S. 1994, Convergence Properties of Cascade Correlation in Function Approximation, Neural Computing & Applications, 2, 142–147.
Drago G.P. and Ridella S. 1996, On the convergence of a growing topology neural algorithms, Neurocomputing, Vol.12, n.2–3, 1996, 171–185.
Fahlman S.E. and Lebiere C. 1990, The Cascade-Correlation Architecture, advances in Information Processing Systems 2, Touretzky, D. (Ed.), 524–532.
Ridella S., Rovetta S., Zunino R., Circular Back Propagation Networks, IEEE Trans. on Neural Networks, Vol. 8, n. 1, January 1997, 84–97.
Zwietering P.J., Aarts E.H.L. and Wessel J., Exact classification with Two-layered perceptrons, International Journal Neural Systems, Vol.3, no.2 (1992), 143–156.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1999 Springer-Verlag London Limited
About this paper
Cite this paper
Drago, G.P., Ridella, S. (1999). Convergence Properties of the Circular Cascade Correlation algorithm. In: Marinaro, M., Tagliaferri, R. (eds) Neural Nets WIRN VIETRI-98. Perspectives in Neural Computing. Springer, London. https://doi.org/10.1007/978-1-4471-0811-5_35
Download citation
DOI: https://doi.org/10.1007/978-1-4471-0811-5_35
Publisher Name: Springer, London
Print ISBN: 978-1-4471-1208-2
Online ISBN: 978-1-4471-0811-5
eBook Packages: Springer Book Archive