User`s guide

8 Model Analysis
This opens an empty plot.
3 In the System Identification Tool window , click each model icon to display
it on the Residual Analysis plot.
Description of the Residual Plot Axes
The top ax
es show the a uto correlation of residuals for the output (whiteness
test). T
he horizontal scale is the number of lags, which is the time difference
(in s amp
les) between the signals at which the correlation is estimated. The
horizon
tal dashed lines on the plot represent the confidence interval of the
corres
ponding e stimates. Any fluctuations within the confidence interval are
consid
ered to be insignificant. T wo of the models,
n4s3 and arx22 3, p roduce
residu
als that enter outside the confidence interval. A good model should have
aresid
ual autocorrelation function within the confidenc e interval, indicatin g
that t
he residuals are uncorrelated.
The bo
ttom axes show the cross-correlation of the residuals with the
input
. A good model should have residuals uncorrelated with past inputs
(inde
pendence test). Evidence of correlation indicates that the model does not
8-22