NGM: Neural Gaussian Mirror for Controlled Feature Selection in Neural Networks.
Xin XingYu GuiChenguang DaiJun S. LiuPublished in: ICMLA (2020)
Keyphrases
- neural network
- feature selection
- network architecture
- pattern recognition
- text categorization
- artificial neural networks
- neural computation
- associative memory
- back propagation
- feature extraction
- information gain
- neural network model
- learning rules
- feature selection algorithms
- recurrent neural networks
- neural learning
- fuzzy logic
- neural model
- nonlinear functions
- multi class
- discriminative features
- feature set
- neural nets
- multilayer perceptron
- field of view
- activation function
- connectionist models
- gaussian mixture model
- rule extraction
- unsupervised learning
- knn
- maximum likelihood
- text classification
- classification accuracy
- feature space
- machine learning
- gaussian noise
- hebbian learning
- support vector machine
- feed forward
- mutual information
- high dimensionality
- decision trees
- gaussian distribution
- dimensionality reduction
- recurrent networks
- fault diagnosis
- self organizing maps
- genetic algorithm ga
- selected features
- radial basis function
- microarray data
- feature subset
- classification models
- multi task