Romain Tavenard

Hi @p3trus This is a good idea indeed. Though, I am not sure when this can be implemented.

Hi @choltz95 100% agree. I also think that more traditional stat models, like linear auto regressive and co would be great to have in future tslearn releases.

Hi @StefanR44 I have this on my side: ```python import numpy import matplotlib.pyplot as plt from tslearn.metrics import sakoe_chiba_mask def plot_constraints(title, sz): for pos in range(sz): plt.plot([-.5, sz - .5],...

If the question is about the kernel itself, I guess both papers by Cuturi are of interest (see for example the dedicated page there: https://marcocuturi.net/GA.html) And if the question is...

Hi @RaffaLinux You are right that soft-dtw cross distances are not parallelized at the moment. Would you be OK to run some tests on your side to check whether it...

Hi @yasirroni I am afraid what you look for is not implemented at the moment. As suggested in #180, if the metrics you want to compare are alignment metrics that...

Hi @enricofallacara Your PR is now in conflict with the master branch since we have refactored the code. I can take care of the merge to resolve the conflicts if...

I have left a few comments. Maybe the main point would be to assess the improvement (in terms of running times) that these changes offer on some benchmark.

Also, tests seem to fail at the moment (let me know if it's due to a bug in my merge).

Hi @GillesVandewiele I'm pretty busy until the end of this week. Will try to catch up early next week. If I don't, do not hesitate to ping me.