This is my first experience in signal processing, so don't treat me badly :) However, I have estimated an impulse-response model using the MATLAB System Identification toolbox. The output, validated on unseen data, has a good fit, approximately around 80%. However, when I export the object imp - its type is idtf - containing the estimated model, I don't know how to evaluate it on unseen data. Obviously, in the structure is contained the polynomial structure of the model, but it bothers me to write a function to estimate it... Is there any quicker procedure?

  • $\begingroup$ Have you tried methods idtf? struct(obj) actually lets you see what is there. $\endgroup$ – user28715 Aug 20 '17 at 17:43

One way of evaluating the model response to arbitrary data is the lsim command. As you don't provide more information with respect to your data and/or model an example below (taken straight from the MATLAB help). See help lsim or doc lsim for more information.

% create model
num = [1 4];
den = [1 20 5];
G = idtf(num,den);

% evaluate response
t = 0:0.01:5;   
u = sin(t);   
y = lsim(G,u,t);

% plot

In addition, if desired, you can evaluate the model impulse, step and/or frequency response using the impulse, step and bode commands.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.