A mutual information based feature selection algorithm
Author:
- 1 October 2011
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 4 (19482914) , 2241-2245
Abstract
The objective of the eliminating process is to reduce the size of the input feature set and at the same time to retain the class discriminatory information. This paper proposes and evaluates a new feature selection algorithm using information theory which is the mutual information (MI) between combinations of input features and the class instead of mutual information between a single input feature and the class for both continuous-valued and discrete-valued features. Comparison studies of new and previously published classification algorithms indicate that the proposed algorithm is robust, stable and efficient.Keywords
This publication has 12 references indexed in Scilit:
- A discrete particle swarm optimization method for feature selection in binary classification problemsEuropean Journal of Operational Research, 2010
- Normalized Mutual Information Feature SelectionIEEE Transactions on Neural Networks, 2009
- A maximum entropy approach to feature selection in knowledge-based authenticationDecision Support Systems, 2008
- Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancyIEEE Transactions on Pattern Analysis and Machine Intelligence, 2005
- Relevant, Irredundant Feature Selection and Noisy Example EliminationIEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2004
- Consistency-based search in feature selectionArtificial Intelligence, 2003
- Feature selection from huge feature sets2002
- A hybrid filter/wrapper approach of feature selection using information theoryPattern Recognition, 2002
- PARTIAL RETRAINING: A NEW APPROACH TO INPUT RELEVANCE DETERMINATIONInternational Journal of Neural Systems, 1999
- Using mutual information for selecting features in supervised neural net learningIEEE Transactions on Neural Networks, 1994