Abstract
A new method to maximize the margin of MLP classifier in classification problems is described. Thismethod is based on a new cost function which minimizes the variance ofthe mean squared error. We show that with this cost function the generalizationperformance increase. This method is tested and compared with the standard mean square errorand is applied to a face detection problem.
Similar content being viewed by others
References
Bernier, O., Collobert, M., Feraud, R., Lemaire, V., Viallet, J. E. and Collobert, D.: MULTRAK: A system for Automatic Multiperson Localization and tracking in real-time, In: International Conference on Image Processing, 1998.
Feraud, R. and Bernier, O.: Ensemble and modular approaches for face detection: a comparison, Neural Information Processing System 10 (1997), 472-478.
Holger, S. and Bengio, Y.: Training method for adaptative boosting of neural networks, Neural Information Processing System, 1998.
Breiman, L.: Bagging predictors, Technical report TR-421, University of California, Berkley, 1994.
Le Cun, Y., Boser, B. and Denker, J. S.: Backpropagation applied to handwritten zip code recognition, Neural Computation 1 (1989), 541-551.
Rosenblatt, M.: Principles of neurodynamics: Perceptron and Theory of Brains mechanisms, Spartan Books, Washington, D.C., 1962.
Rumelhart, D. E., Hinton, G. E. and Williams, R. J.: Learning internal representations by error propagation, Parallel Distributed Processing: Explorations in the Microstructures of Cognition 1 (1986), 318-362.
Schapire, R. E., Freund, Y., Bartlett, P. and Lee,W. S.: Boosting the margin: A new explanation for the effectiveness of voting methods, Machines That Learn, 1997.
White, H. and Hornik, K.: Mutilayer feedforward networks are universal approximators, Neural Network 2 (1989), 359-366.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Lemaire, V., Bernier, O., Collobert, D. et al. A New Method to Increase the Margin of Multilayer Perceptrons. Neural Processing Letters 11, 7–15 (2000). https://doi.org/10.1023/A:1009607527399
Issue Date:
DOI: https://doi.org/10.1023/A:1009607527399