Abstract
The tri-state neuron is introduced to simulate the tri-state property of neurons in the nervous system. The double sigmoid function is utilized as the activation function of the tri-state neural network. The convergence rate and the generalization ability of the tri-state network are compared with those of the traditional bi-state network with single sigmoid function. It is shown analytically and computationally that (1) the tri-state network has a faster convergence rate than the bi-state network and that (2) the generalization ability of these two networks are comparable.
Original language | English (US) |
---|---|
Pages (from-to) | 235-246 |
Number of pages | 12 |
Journal | Intelligent Automation and Soft Computing |
Volume | 1 |
Issue number | 3 |
DOIs | |
State | Published - 1995 |
Bibliographical note
Funding Information:The authors would like to acknowledge the valuable contribution of Bo Xu at various stages of this research project. Support of this work by IUPUI Research Investment Fund is gratefully acknowledged. We would also like to acknowledge the useful suggestions of one of the anonymous reviewers.
Keywords
- Double sigmoid
- Fast convergence
- Prediction
- Tri-state neuron
- Volterra time series