1
$\begingroup$

I have to compare pairs of audio strems as 1d time series. Looking at the aligned trajectories I need to either cluster them together or assume they arise from independent generators. I remember hearing of Dirichlet processes being applied. Could try to defined a hidden Markov model on the structure, not sure yet how the emission probabilities would work out, but it must be possible.

any recommendations? Best

EDITS- set of 1D data time series s in {1...N}, another set of trajectories q in {1...N} all over a discrete time domain t=0...t=T under a uniform interval. Task, determine whether under certain assumptions (eg. iid, normality etc), is the generator under these assumptions a valid null hypothesis to have generated the two categories of trajectories s and q? Or is the null hypothesis abandoned for a more complex model?

So do we use an autoregressive model? a Gaussian process would be hard to interpret the inverse covariance. Some parametric model would seem to be valid. Don't know... any ideas?

$\endgroup$

closed as not a real question by russellpierce, gung, whuber Feb 13 '13 at 15:36

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ I think it highly unlikely that you will get answered without clarifying this question. $\endgroup$ – Shane Aug 27 '10 at 13:48
  • $\begingroup$ Cautious +1 to the edit. In truth, I think the question is still not perfectly clear, but I know I'd love to read the answers! $\endgroup$ – walkytalky Aug 27 '10 at 23:14
  • $\begingroup$ Do I understand it right that you have two sets of time series, and you want to check whether or not they differ significantly? $\endgroup$ – Joris Meys Sep 2 '10 at 13:59
  • $\begingroup$ @Joris Meys- exactly $\endgroup$ – Vass Sep 8 '10 at 10:25
  • $\begingroup$ Are you trying to decide if a pair of voice recordings is stereo recording of the same sound source or coming from two independent sources? $\endgroup$ – GaBorgulya Apr 3 '11 at 22:32
1
$\begingroup$

To answer this question it needs to be clear along which dimensions you want to perform your categorization. There are many possible answers in this regard. For example: you could conduct an FFT on each dataset and then compare the results. If they are remarkably similar across a number of parameters (up to you to judge) then perhaps they can be considered to be categorically similar. However, without further information from you, this question will be impossible to answer.

$\endgroup$

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