getEuclideanDistance should depends on the state of timeseries about sfa HOT 4 OPEN

Thank you for pointing this out.

There are two cases:

1. the query is z-normed: thus we have to z-norm all the subsequences.

2. query is not z-normed:
   a) we can either force the query to be z-normed, which is one call to query.norm() before performing the query or
   b) we can not apply z-normalization at all, which would be a check in getDistance.

Typically, we force the query to be z-normed in this case. However, I agree that it is possible to leave out z-norm, too. The proposed pull-request however, breaks the code by removing z-norm for subsequences.

from sfa.

For my understanding, it's better to return the un-normalized distance measure. So a potential fix will be: if the query is normalized -> unnormalize its values, if the timeseries is normalized -> un-normalize the values -> then calculate the euclidean distance between original un-normalized versions of both timeseries data and query :)
Big thanks anyway!

from sfa.

Removing it, is not sooo easy. For example, SFA uses the momentary Fourier transform. The MFT applies z-normalization to each subsequence:

The same does TimeSeries.getDisjointSequences and TimeSeries.getSubsequences

So, when removing it from getDistance, you suddenly try to lower bound un-normalized subsequences with z-normalized words. This should fail.

from sfa.

yes i know, that's why i didn't propose a fix. I saw that in subsequence matching, each subsequence is normalized with different mean and std, so i can't un-normalized. maybe we need 2 separate methods.

from sfa.