If I have two completely different time series at equally sampled times, but the length of each of these series is different. How do I statistically compare them? Dynamic Time Warping?

Also, as a follow up question. How do I go about this if each time series is of say 12 dimensions?

  • $\begingroup$ the question is two broad. what exactly are you trying to compare? which features? you can always start with spectral analysis, of course. $\endgroup$ – Aksakal Feb 27 '14 at 18:33
  • $\begingroup$ I've extracted 22 or so features from audio tracks of varying length and I'd like to compare each time series of audio features with another. $\endgroup$ – daithioronain Feb 27 '14 at 18:37
  • $\begingroup$ are you building Shazam? $\endgroup$ – Aksakal Feb 27 '14 at 18:38
  • $\begingroup$ Ha, if only! No it's for an automatic mixing PhD. $\endgroup$ – daithioronain Feb 27 '14 at 18:39
  • $\begingroup$ Each of my audio track would be just one instrument playing $\endgroup$ – daithioronain Feb 27 '14 at 18:40

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.