Back to Search
Start Over
Selecting features for neural network committees
- Source :
- Proceedings of the 2002 International Joint Conference on Neural Networks. IJCNN'02 (Cat. No.02CH37290).
- Publication Year :
- 2003
- Publisher :
- IEEE, 2003.
-
Abstract
- We present a neural network based approach for identifying salient features for classification in neural networks. Our approach involves neural network training with an augmented cross-entropy error function. The augmented error function forces the neural network to keep low derivatives of the transfer functions of neurons when learning a classification task. Such an approach reduces the output sensitivity to input changes. Feature selection is based on the reaction of the cross-validation data set classification error due to the removal of the individual features. We compared the approach with two other neural network based feature selection methods. The algorithm developed outperforms the methods by achieved a higher classification accuracy on three real world problems tested.
- Subjects :
- Quantitative Biology::Neurons and Cognition
Artificial neural network
Time delay neural network
Computer science
Entropy (statistical thermodynamics)
business.industry
Deep learning
Feature extraction
Feature selection
Pattern recognition
Machine learning
computer.software_genre
Probabilistic neural network
Entropy (classical thermodynamics)
Error function
Entropy (information theory)
Artificial intelligence
Entropy (energy dispersal)
Stochastic neural network
business
computer
Subjects
Details
- Database :
- OpenAIRE
- Journal :
- Proceedings of the 2002 International Joint Conference on Neural Networks. IJCNN'02 (Cat. No.02CH37290)
- Accession number :
- edsair.doi...........1611cdd637da1efb06eaf42ca44b8cb7