Abstract
When training set is unbalanced, the conventional least square error (LSE) training strategy is less efficient to train neural network (NN) for classification because it often lead the NN to overcompensate for the dominant group. Therefore, in this paper a dynamic threshold learning algorithm (DTLA) is proposed as the substitute for the conventional LSE algorithm. This method uses multiple dynamic threshold parameters to gradually remove some training patterns that can be classified correctly by current Radial Basis Function (RBF) network out of the training set during training process, which changes the unbalanced training problem into a balanced training problem and improves the classification rate of the small group. Moreover, we use the dynamical threshold learning algorithm to classify the remote sensing images, when the unbalanced level of classes is high, a good effect is obtained.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Murphey, Y.L., Guo, H., Feldkamp, L.A.: Neural Learning from Unbalanced Data. Applied Intelligence 21(2), 117–128 (2004)
David, L., Andrew, R.W.: Optimized Feature Extraction and the Bayes Decision in Feed-Forward Classifier Networks. IEEE Trans. on PAMI 13(4), 355–364 (1991)
Friedhelm, S., Kestler, H.A., Palm, G.: Three Learning Phases for Radial-Basis-Function Networks. Neural Networks 14(4-5), 439–458 (2001)
Sontag, E., Sussman, H.: Backpropagation Separates When Perceptrons Do. In: Proceeding IEEE International Joint Conference on Neural Networks (Washington DC), vol. 1, pp. 639–642 (1989)
Francesco, L., Marco, S.: Efficient Training of RBF Neural Networks for Pattern Recognition. IEEE Trans on Neural Networks 12(5), 1235–1241 (2001)
Sami, M.A., Linwood, J., Park, J.D., Shannon, M.F.: A Neural Network Algorithm for Sea Ice Edge Classification. IEEE Trans. on Geosciene and Remote Sensing 35(4), 817–826 (1997)
James, J.S., Timothy, J.M.: A Recurrent Neural Network Classifier for Improved Retrievals of Areal Extent of Snow Cover. IEEE Trans. on Geoscience and Remote Sensing 39(10), 2135–2147 (2001)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Li, BY., Peng, J., Chen, YQ., Jin, YQ. (2006). Classifying Unbalanced Pattern Groups by Training Neural Network. In: Wang, J., Yi, Z., Zurada, J.M., Lu, BL., Yin, H. (eds) Advances in Neural Networks - ISNN 2006. ISNN 2006. Lecture Notes in Computer Science, vol 3972. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11760023_2
Download citation
DOI: https://doi.org/10.1007/11760023_2
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-34437-7
Online ISBN: 978-3-540-34438-4
eBook Packages: Computer ScienceComputer Science (R0)