Abstract
This paper presents a new algorithm for designing neural network ensembles for classification problems with noise. The idea behind this new algorithm is to encourage different individual networks in an ensemble to learn different parts or aspects of the training data so that the whole ensemble can learn the whole training data better. Negatively correlated neural networks are trained with a novel correlation penalty term in the error function to encourage such specialization. In our algorithm, individual networks are trained simultaneously rather than independently or sequentially. This provides an opportunity for different networks to interact with each other and to specialize. Experiments on two real-world problems demonstrate that the new algorithm can produce neural network ensembles with good generalization ability.
Similar content being viewed by others
References
Liu Y, Yao X (1998) A cooperative ensemble learning system. In: Proceedings of the 1998 IEEE International Joint Conference on Neural Networks (IJCNN'98), Anchorage, USA, May 4–9, 1998, pp 2202–2207
Sharkey AJC (1996) On combining artificial neural nets. Connect Sci 8:299–313
Yao X, Liu Y (1996) Ensemble structure of evolutionary artificial neural networks. In: Proceedings of the 1996 IEEE International Conference on Evolutionary Computation (ICEC'96), IEEE Press, pp 659–664
White H (1989) Learning in artificial neural networks: a statistical perspective. Neural Comput 1:425–464
Haykin S (1994) Neural networks: a comprehensive foundation. Macmillan College Publishing, pp 151–152
Yao X, Liu Y (1997) A new evolutionary system for evolving artificial neural networks. IEEE Trans Neural Networks 8:694–713
Yao X, Liu Y (1998) Making use of population information in evolutionary artificial neural networks. IEEE Trans Syst, Man, Cybern, Part B, Cybernetics 28B(3):417–425
Michie D, Spiegelhalter DJ, Taylor CC (1994) Machine learning, neural and statistical classification. Ellis Horwood, London
Setiono R, Hui LCK (1995) Use of a quasi-Newton method in a feedforward neural network construction algorithm. IEEE Trans Neural Networks 6:273–277
Prechelt L (1994) Proben 1—a set of neural network benchmark problems and benchmarking rules. Fakultät für Informatik, Universität Karlsruhe, 76128 Karlsruhe, Germany, 21/94, September
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Liu, Y., Yao, X. Negatively correlated neural networks for classification. Artif Life Robotics 3, 255–259 (1999). https://doi.org/10.1007/BF02481190
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF02481190