Why does prediction by AR model cause a time lag?

Please tell me why theoretically.

  • 2
    $\begingroup$ in AR model you use previous values of signal to predict the future values so to do a prediction you have to know some previous values of the signal depending on the order of model. Generally the time lag regarded as the number of samples you need to know and it's not the lag of output signal. $\endgroup$ – Mohammad M Aug 21 '17 at 11:47
  • $\begingroup$ What Mohammad said! If you read the wikipedia article on AR models, you'll find that "this is basically equivalent to an … IIR filter": So, the analysis methods for group delay of IIR filters apply. That's a rich field of theory, and you can find plenty about "IIR filter delay" online. $\endgroup$ – Marcus Müller Aug 21 '17 at 17:53
  • $\begingroup$ I agree with the other comments but try looking at it this way: you are basing your estimate/output on the history of the data so that time has to have passed. $\endgroup$ – rrogers Aug 23 '17 at 12:57

Your Answer

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

Browse other questions tagged or ask your own question.