Maximization of Mutual Information for Supervised Linear Feature Extraction
- 4 September 2007
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks
- Vol. 18 (5), 1433-1441
- https://doi.org/10.1109/tnn.2007.891630
Abstract
In this paper, we present a novel scheme for linear feature extraction in classification. The method is based on the maximization of the mutual information (MI) between the features extracted and the classes. The sum of the MI corresponding to each of the features is taken as an heuristic that approximates the MI of the whole output vector. Then, a component-by-component gradient-ascent method is proposed for the maximization of the MI, similar to the gradient-based entropy optimization used in independent component analysis (ICA). The simulation results show that not only is the method competitive when compared to existing supervised feature extraction methods in all cases studied, but it also remarkably outperform them when the data are characterized by strongly nonlinear boundaries between classes.Keywords
This publication has 12 references indexed in Scilit:
- A Fixed-Point Algorithm for Finding the Optimal Covariance Matrix in Kernel Density ModelingPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2006
- Discriminative Components of DataIEEE Transactions on Neural Networks, 2005
- Estimating mutual informationPhysical Review E, 2004
- Analyzing Neural Responses to Natural Signals: Maximally Informative DimensionsNeural Computation, 2004
- 10.1162/153244303322753742Applied Physics Letters, 2000
- Blind signal separation: statistical principlesProceedings of the IEEE, 1998
- Nonlinear Component Analysis as a Kernel Eigenvalue ProblemNeural Computation, 1998
- Using mutual information for selecting features in supervised neural net learningIEEE Transactions on Neural Networks, 1994
- Sliced Inverse Regression for Dimension ReductionJournal of the American Statistical Association, 1991
- Sliced Inverse Regression for Dimension ReductionJournal of the American Statistical Association, 1991