L. K. Hansen, L. Liisberg, and P. Salamon, Ensemble methods for handwritten digit recognition, in: Proceedings of the IEEE Workshop on Neural Networks for Signal Processing, Copenhagen, Denmark, 1992, pp. 333-342.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....i.e. training multiple artificial neural networks and combining their predictions. Subsequently there appears a hot wave in investigating artificial neural network ensembles [27] and this technique has already been successfully applied to diverse domains such as optical character recognition [7] [ 11 ], 17] face recognition [10] 14] scientific image analysis [4] seismic signals classification [29] etc. In general, an artificial neural network ensemble is built in two steps, that is, generating component artificial neural networks and then combining their predictions. As for generating ....

L. K. Hansen, L. Liisberg, and P. Salamon, "Ensemble methods for handwritten digit recognition," in Proc. IEEE tVorkshop on Neural Networks for Signal Processing, Helsingoer, Denmark, 1992, pp.333- 342.

....is easy to use, neural network ensemble is regarded as a promising methodology that can profit not only experts in neural computing but also ordinary engineers in realworld applications. And neural network ensemble has already been used in many real domains such as handwritten digit recognition [Hansen et al. 1992], scientific image analysis [Cherkauer, 1996] face recognition [Gutta and Wechsler, 1996; Huang et al. 2000] OCR [Mao, 1998] seismic signals classification [Shimshoni and Intrator, 1998] etc. Many works have been done in investigating why and how neural network ensemble works. The classical ....

L. K. Hansen, L. Liisberg L, and P. Salamon. Ensemble methods for handwritten digit recognition. In Proceedings of the IEEE-SP Workshop on Neural Networks for Signal Processing, pages 333342, 1992. IEEE Computer Society.

....remarkably well and is vey easy to use, neural network ensemble is regarded as a promising methodology that can benefit not only experts in neural computing but also ordinary engineers. And neural network ensemble has already been used in many real domains such as handwritten digit recognition[3], scientific image analysis[4] face recognition[5] 6] OCR[7] seismic signal classification[8] etc. Many works have been done to investigat why and how neural network ensemble works. The classical one is Krogh and Vedelsby[9] s work, in which they derived the famous equation E = E A . It ....

L. K. Hansen, L. Liisberg, and P. Salamon, "Ensemble methods for handwritten digit recognition," In Proc. IEEE-SP Workshop on Neural Networks for Signal Processing, pp. 333342, 1992, IEEE Computer Society.

.... ensemble trains a collection of a finite number of neural networks for the same task [21] Since it could significantly improve the generalization ability of learning systems, it has been successfully applied to diversified domains such as face recognition [9, 14] handwritten digit recognition [10], optical character recognition [7, 16] scientific image analysis [4] medical diagnosis [5, 22] seismic signals classification [20] etc. It has been found that the combination of neural networks and rule learning is powerful in some cases [1] Considering that neural network ensemble is ....

L. K. Hansen, L. Liisberg, and P. Salamon, "Ensemble methods for handwritten digit recognition", In Proc. IEEE Workshop on Neural Networks for Signal Processing, Helsingoer, Denmark, pp.333-342, 1992.

....of neural networks, i.e. training many neural networks and then combining their predictions. In recent years, neural network ensemble has become a very hot topic and has already been successfully applied to diverse real domains such as face recognition [13, 17] handwritten digit recognition [14], optical character recognition [8, 19] scientific image analysis [5] medical diagnosis [6, 26] seismic signals classification [23] etc. Recently, Zhou et al. 25] showed for neural regressors that output continuous values, ensembling an appropriate subset of individual neural networks is ....

L. K. Hansen, L. Liisberg, and P. Salamon, "Ensemble methods for handwritten digit recognition", In Proc. IEEE Workshop on Neural Networks for Signal Processing, Helsingoer, Denmark, pp.333-342, 1992.

....through training several neural classifiers and then combining their predictions via voting. Since neural network ensemble has great potential, many researchers run into this area and this technology has already been successfully applied to many domains such as optical character recognition [17, 26], face recognition [14, 18] scientific image analysis [6] medical diagnosis [48] seismic signals classification [39] etc. In general, a neural network ensemble is built in two stages, i.e. training several individual neural networks and then combining their predictions. As for training ....

L. K. Hansen, L. Liisberg L, and P. Salamon, Ensemble methods for handwritten digit recognition, in: Proceedings of the IEEE Workshop on Neural Networks for Signal Processing, Copenhagen, Denmark, 1992, pp. 333-342.

....of the 1990s, Hansen and Salamon [9] showed that the generalization ability of an artificial neural network system can be significantly improved through ensembling artificial neural networks, i.e. training several artificial neural networks and combining their predictions. Later, Hansen et al. [10] applied artificial neural network ensemble to handwritten digit recognition and attained astonishing good results whose accuracy is 20 25 better than that of the best individual artificial neural network. Subsequently there appears a hot wave in investigating artificial neural network ensembles, ....

....neural network ensembles work remarkably well and are easy to be used, they are regarded as a promising methodology that can profit not only experts in artificial neural network research but also engineers in real world applications. Besides Hansen et al. s work in handwritten digit recognition [10], artificial neural network ensembles have already been applied to many real world domains such as scientific image analysis [4] face recognition [8, 13] OCR [19] seismic signals classification [26] breast cancer diagnosis [25] and in vitro fertilization treatment [7] 3. Lung cancer ....

Hansen LK, Liisberg L, Salamon P. Ensemble methods for handwritten digit recognition. In: Proceedings of the IEEE-SP Workshop on Neural Networks for Signal Processing, 1992. p.333-342.

....Keywords: Feature assessment; Feature selection; Neural networks; Bayes classifier; Pruning; Insight; Feature metrics; Feature measures 1. Introduction Multi layer perceptrons (MLPs) have been trained to perform various classification tasks (Cibas et al. 1996; Cunningham et al. 1992; Hansen et al. 1992; Harrison et al. 1991) Hart et al. 1989; Hripcsak, 1990; Moallemi, 1991; Poli et al. 1991; Schiler et al. 1992; Schizas et al. 1990; Vogelsang, 1993) An MLP classifier performs a mapping from an input (feature or attribute) space onto an output (class) space. Cases are represented in the ....

Hansen, L. K., Liisberg, C., & Salamon, P. (1992). Ensemble methods for handwritten digit recognition. In S. Y. Kung, F. Fallside, J. A. Sorenson, & C. A. Kaufmann (Eds.), Proceedings of the 1992 IEEE workshop on neural networks for signal processing (pp. 333--342). NJ: IEEE.

....by the data and the estimates will no longer be accurate. We must either have a method for stopping the increase in complexity (e.g. the standard approach is cross validation) or we must increase the data set fast enough as the complexity increases. 1.3. 8 Hansen s Ensemble Performance Estimate Hansen et al. 1992,1990) develop a theoretical result for a special case of averaging which they call a plurality decision. This is a majority rule where each estimator has an equal vote during classification. 8 Starting with a population of N classifiers voting for M classes, Hansen assumes that all of the ....

....estimator was not only better than the population average but it was also as good as or better than the naive estimator in every case. These results for difficult, real world classification tasks show that the BEM estimator is 33 EXPERIMENT DATA SET TEST FOM Perrone (1992) NIST Numerals 94.7 Hansen et al. 1992) NIST Numerals 83.0 Drucker et al. 1992) NIST Numerals 92.0 Scofield et al. 1991) NIST Numerals 89.7 Xu et al. 1992) USPS Numerals 95.0 Denker et al. 1989) USPS Numerals 76.0 Le Cun et al. 1990,1989) USPS Numerals 81.0 Fontaine Shastri (1992) USPS Numerals 75.4 Drucker et al. 1992) ....

Hansen, L. K., Liisberg, C., and Salamon, P. (1992). Ensemble methods for handwritten digit recognition. In Kung, S. Y., Fallside, F., and Kamm, C. A., editors, Neural Networks for Signal Processing II: Proceedings of the 1992 IEEE Workshop, pages 333--342. IEEE.

No context found.

L. K. Hansen, L. Liisberg, and P. Salamon, Ensemble methods for handwritten digit recognition, in: Proceedings of the IEEE Workshop on Neural Networks for Signal Processing, Copenhagen, Denmark, 1992, pp. 333-342.

No context found.

Hansen L K, Liisberg C, Salamon P., "Ensemble Methods for Handwritten Digit Recognition," Proceedings of the 1992.

No context found.

L. K. Hansen, L. Liisberg, and P. Salamon, Ensemble methods for handwritten digit recognition, in: Proceedings of the IEEE Workshop on Neural Networks for Signal Processing, Copenhagen, Denmark, 1992, pp. 333-342.

No context found.

L. K. Hansen, L. Liisberg L, and P. Salamon, Ensemble methods for handwritten digit recognition, in Proc. IEEE Workshop on Neural Networks for Signal Processing, Copenhagen, Denmark, 1992, 333-342.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC