Abstract:
The well-known methods of choosing the process model structure are analyzed when the model is restored from noisy experimental data and its assignment implies forecasting the process output response as a function of the input signals. The structural identification methods are analyzed with the model being designed in two stages, choiceof the structure class and parametric identification. Special attention is given to choice? of the model structure for linear dynamic processes.
Fulltext:
PDF file (4568 kB)