Identification Procedure

Step-1. Experiment

Apply input and observe the effect on output results that is used to infer a model of the system.

Step-2. Model Structure Selection

After doing experiment, select set of candidate model. There are two aspects

• Select a group of appropriate model structure. example, linear, multilayer perceptron, wavelets, etc.
• Select a subset of models.

Step-3. Estimate Model

To pick one particular model based on some performance criterion.

Step-4. Validate Model

To check whether the estimated model meets the requirements. If yes, then this is the required model for the investigating system otherwise go to steps 1 or 2 or 3 and make changes till the model meets the requirements.