Avoiding overfitting of multilayer perceptrons by training derivatives.
Vsevolod I. AvrutskiyPublished in: CoRR (2018)
Keyphrases
- multilayer perceptron
- backpropagation algorithm
- radial basis function network
- feedforward neural networks
- back propagation
- neural network
- hidden layer
- multilayer neural network
- radial basis function
- training error
- training algorithm
- feed forward neural networks
- artificial neural networks
- avoid overfitting
- cross validation
- number of hidden neurons
- rbf network
- commonly applied
- neuro fuzzy models
- training phase
- multi layer perceptron
- neural network model
- training set
- activation function
- bp algorithm
- text classification
- single hidden layer
- pattern recognition
- decision trees
- machine learning