A Pipeline Approach in Identifying Important Input Features from Neural Networks.
Yuyu HeChih LaiDalma Martinovic-WeigeltZezheng LongPublished in: SoSE (2019)
Keyphrases
- input features
- neural network
- activation function
- machine learning algorithms
- feature values
- artificial neural networks
- classification algorithm
- pattern recognition
- feature selection
- hidden layer
- class labels
- back propagation
- basis functions
- neural network model
- feed forward
- support vector machine
- fuzzy logic
- labeled examples
- training data
- fault diagnosis
- cross validation
- semi supervised learning
- recurrent neural networks
- learning rate
- network architecture
- multi layer perceptron
- similarity measure