3

While thinking about similarity between two time series, one can use DTW to approach the issue. There is a Python package for that mlpy.

It is also said to compare time series via simple euclidean distance.

Is there a Python package to this?

Community
  • 1
KcFnMi
  • 343
  • 1
  • 4
  • 8
  • Google: "Euclidean Distance python" results in http://stackoverflow.com/questions/1401712/how-can-the-euclidean-distance-be-calculated-with-numpy and in http://www.codehamster.com/2015/03/09/different-ways-to-calculate-the-euclidean-distance-in-python/ – Nikolas Rieble Jan 05 '17 at 08:54
  • Start by telling us what time-intervals the two series actually represent? And have same length? – smci Mar 05 '17 at 02:37

1 Answers1

6

This only makes sense if you look at the time series as vectors with same length. In that case, there is no need for a package. Just use, for example, numpy.linalg.norm(a-b).

oW_
  • 6,347
  • 4
  • 28
  • 47