Training Auto-encoders Effectively via Eliminating Task-irrelevant Input Variables.
Hui ShenDehua LiHong WuZhaoxiang ZangPublished in: CoRR (2016)
Keyphrases
- input variables
- variable selection
- artificial neural networks
- neural network model
- membership functions
- fuzzy rules
- number of input variables
- hidden neurons
- supervised learning
- reduced set
- training phase
- neural network
- fuzzy sets
- multi layer perceptron
- ann models
- principal component analysis
- cross validation
- input image
- high dimensional
- feature selection
- genetic algorithm
- real world
- ls svm