system identification gui - fit to estimation data

Hi. i was studying the system identification gui help at link below: http://www.mathworks.com/help/ident/gs/identify-linear-models-using-the-gui.html

after computing amx3322 with the fitness of 89.6 , the help says: "To view parameter uncertainties, click Present in the Data/model Info dialog box, and view the model information in the MATLAB Command Window." at the end lines you see this sentence:"Fit to estimation data: 95.3%" what is the meaning of this line; we have computed the fitness to 89.6 but it says 95.3 is it something diffrent? and i don't understand the meaning of this one:FPE: 0.00163, MSE: 0.00155;

Answer 1

[Rajiv Singh]

When the amx3322 model was estimated, it was estimated to minimize the 1-step ahead prediction error. The fit value between the 1-steap ahead predicted response of this model to measured data was 95.3%. The Model Output plot shows (by default) the fit between the simulated response of the model and the measured data. This, in general, is different from the predicted response. To see the 95.5% fit in Model Output, use the Options menu of the plot. First set the prediction horizon to 1 and then check the option to perform 1-step ahead prediction rather than simulation. Note that simulation means infinite horizon prediction.

Fit value compute is the normalized root mean square (NRMSE) value. FPE refers to final prediction error and MSE is mean square error. All are different measures (along with also AIC) of model quality. See:

http://www.mathworks.com/help/ident/ref/goodnessoffit.html

http://www.mathworks.com/help/ident/ref/aic.html

http://www.mathworks.com/help/ident/ref/fpe.html

[garima sharma]

hello everyone,

what is the meaning of negative value of Fit to estimtion (e.g. Fit to estimation data = -1.5*10^23%) although i seleted order based on model order selection?