A non-iterative method for pruning hidden neurons in neural networks with random weights.
Pablo A. HenríquezGonzalo A. RuzPublished in: Appl. Soft Comput. (2018)
Keyphrases
- hidden neurons
- neural network
- activation function
- artificial neural networks
- feedforward neural networks
- generalization ability
- input variables
- hidden layer
- training speed
- back propagation
- single hidden layer
- neural network model
- multilayer perceptron
- feed forward
- fault diagnosis
- recurrent neural networks
- training algorithm
- pattern recognition
- nonlinear functions
- bp neural network
- model selection
- particle swarm optimization
- feature vectors
- multi layer perceptron
- support vector
- connection weights
- learning algorithm
- genetic algorithm
- machine learning