[flashfaffe2]

Maybe this is off topic but how this algorithm adapt itself to time series?

I mean you want to cluster data which might be autocorrelated... Does this case is taken into account or do you need to preprocess your data beforehand?