I have two continuous time series of discrete events. The data from them are timestamps of occurences, e.g.

X = ['13:28', '18:35', '19:22', '19:58']
Y = ['14:28', '15:35', '19:11']

Basic simple approach:

  1. Discretize the occurrences (e.g. indicating if during a specific hour, the event happened).

  2. Apply Jaccard on the discretized data.

I am wondering if there is any more sophisticated approach. The major problem is the information loss at discretization.

  • $\begingroup$ Pls tell us what's your objective, how are you going to use this correlation. $\endgroup$ – horaceT Oct 13 '17 at 16:48
  • $\begingroup$ Simply to decide whether to consider X events as happening at the same time as Y events. A tolerance parameter can be considered, e.g. plus minus 10 minutes. $\endgroup$ – Karel Macek Oct 13 '17 at 19:08

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.