User`s guide

Using Residual Analysis Plots to Validate M odels

• For output-error (OE) models and when using instrumental-variable (IV)

methods, make sure that your model shows independence of

e and u,and

pay less a ttention to the results of the whiteness of

In this case, the modeling focus is on the dynamics G and not the

disturbance properties H.

• Correlation between residuals and input for negative lags, is not necessarily

an indication of an inaccurate model.

When current residuals at time t affect future input values, there might

be feedback in your system. In the case of feedback, concentrate on the

positive lags in the cross-correlation plot during model validation.

Supported Model Types

You can validate parametric linear and nonlinear models by checking the

behavior of the model residuals. For a description of re sidual a nalysis, see

“What Does the Residuals Plot Show?” on page 8-17.

Note For nonparametric models, including impulse-response, step-response,

and frequency-response models, residual analysis plots are not available. For

time-series m odels , you can only generate model-output plots for parametric

models using time-domain time-serie s (no input) measured d ata.

What Does the Residuals Plot Show?

Residual analysis plots sho w different information depending on whether you

use time-domain or frequency-doma in input-output validation data.

For tim e-domain validation data, the plot shows the following two axes:

• Auto correlation function of the residuals for each output

• Cross-correlation between the input and the residuals for each input-output

pair

Note For time-series m odels , the re sidual analysis plot does not provide

any input-residual correlation plots.

8-17