Nonlinear feature transforms using maximum mutual information
- 13 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 4 (10987576) , 2756-2761 vol.4
- https://doi.org/10.1109/ijcnn.2001.938809
Abstract
Finding the right features is an essential part of a pattern recognition system. This can be accomplished either by selection or by a transform from a larger number of "raw" features. In this work we learn nonlinear dimension reducing discriminative transforms that are implemented as neural networks, either as radial basis function networks or as multilayer perceptrons. As the criterion, we use the joint mutual information (MI) between the class labels of training data and transformed features. Our measure of MI makes use of Renyi entropy as formulated by Principe et al. (1998, 2000). Resulting low-dimensional features enable a classifier to operate with less computational resources and memory without compromising the accuracy.Keywords
This publication has 10 references indexed in Scilit:
- A methodology for information theoretic feature extractionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Pose estimation in SAR using an information theoretic criterionPublished by SPIE-Intl Soc Optical Eng ,1998
- Feature Selection for Knowledge Discovery and Data MiningPublished by Springer Nature ,1998
- Bhattacharyya distance feature selectionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1996
- Nonlinear feature extraction with radial basis functions using a weighted multidimensional scaling stress measurePublished by Institute of Electrical and Electronics Engineers (IEEE) ,1996
- Neural Networks for Pattern RecognitionPublished by Oxford University Press (OUP) ,1995
- Artificial neural networks for feature extraction and multivariate data projectionIEEE Transactions on Neural Networks, 1995
- Using mutual information for selecting features in supervised neural net learningIEEE Transactions on Neural Networks, 1994
- Nonlinear principal component analysis using autoassociative neural networksAIChE Journal, 1991
- An optimal orthonormal system for discriminant analysisPattern Recognition, 1985