Reducing noise impact on MLP training - Techniques and algorithms to provide noise-robustness in MLP network training.
Miroslaw KordosAndrzej RusieckiPublished in: Soft Comput. (2016)
Keyphrases
- multi layer perceptron
- radial basis function network
- multilayer perceptron
- neural network
- noisy data
- greater robustness
- artificial neural networks
- learning algorithm
- computational efficiency
- computational complexity
- output layer
- signal to noise ratio
- data structure
- training examples
- training and testing data
- early stopping
- recurrent networks
- data sets
- missing data
- multilayer neural network
- neural classifier
- hidden neurons
- random noise
- noise model
- hidden layer
- radial basis function
- training phase
- gaussian noise
- input data
- median filter
- feedforward neural networks
- training process
- optimization problems
- noise level
- training data
- online learning
- bayesian networks
- network structure
- computationally efficient