Abstract
The performance of the neural networks estimated with Bayesian regularization to model and forecast time series was studied where the data generating process was linear. The technique of Bayesian regularization was designed to avoid overfitting to obtain balance between the number of parameters. Forecasts with larger RMSE and MAE statistics were produced in small samples, neural networks.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 237-242 |
| Number of pages | 6 |
| Journal | International Journal of Engineering Intelligent Systems for Electrical Engineering and Communications |
| Volume | 9 |
| Issue number | 4 |
| State | Published - Dec 2001 |
| Externally published | Yes |
Keywords
- Bayesian regularization
- Neural networks
- Nonlinear forecasting
- Nonlinear time series
ASJC Scopus subject areas
- Control and Systems Engineering
- Computer Science Applications
- Electrical and Electronic Engineering