You should use different data sets to estimate and validate your model for best validation results.
Select portions of data for identification in the app or at the command line.
You can create a new data set in the System Identification app by extracting subsets of input and output channels from an existing data set.
You can create a time-domain or frequency-domain data set in the System Identification app that includes several experiments. Identifying models for multiexperiment data results in an average model.
You can create
iddata objects that contain several experiments.
Identifying models for an
iddata object with multiple experiments
results in an average model.
This example shows how to deal with multiple experiments and merging models when working with System Identification Toolbox™ for estimating and refining models.