I have a collection of data points. Each point has 6 dimensions (x1, x2,...x6). I want to find a relation between two dimension (e.g. x1 vs x2). What I have been doing so far is look for points where the other dimensions (x3 to x6) are relatively constant, by defining a band. This way I would get several groups of data points where only the two dimensions of interest would change.

I was wondering if there is a better way of analyzing the relationship between these two dimensions. I looked at PCA, but I have a feeling that it does not help me much. If I reduce the problem to two dimensions the axes are basically meaningless.

Can you guys give me some directions to look at?


You might take a look at canonical correlation analysis. It tries to find correlation between two sets of data. I guess you could adapt it to explore your data.

  • $\begingroup$ Yes, I just edited it, thx. $\endgroup$ – Robin Feb 24 '17 at 9:12
  • $\begingroup$ @debzsud You corrected the wrong "two", I fixed it all, and added link to wikipedia article. (my edit is pending approval at this moment) $\endgroup$ – VividD May 25 '17 at 19:52

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.