Adaptative time constants improve the prediction capability of recurrent neural networks.
Jean-Philippe DrayeDavor PavisicGuy CheronGaetan LibertPublished in: Neural Process. Lett. (1995)
Keyphrases
- recurrent neural networks
- feed forward
- neural network
- recurrent networks
- reservoir computing
- chaotic time series
- artificial neural networks
- prediction accuracy
- echo state networks
- neural model
- complex valued
- feedforward neural networks
- predictive clustering trees
- cascade correlation
- hebbian learning
- mathematical model
- long term
- artificial intelligence
- machine learning