Different matrix profiles between tsmp & STUMP

[snvv]

Hello and thank you for your excellent software.

I tried to reproduce the MP of a time series (ts) using stumby and tsmp.
ts: 0 1 3 2 9 1 14 15 1 2 2 10 7
tsmp mp: 0.6 0.3 1.6 1.1 1.3 1.8 3.0 2.8 0.3 0.6
stumby mp: 6.9 1.4 6.2 7.9 11.4 13.6 14.1 14.0 1.4 6.2
I think the difference should be due to some normalization that takes place in tsmp that is hidden of the end user. Is it possible to pass a normalization parameter to tsmp to allow the end user to choose the normalization method or to use unnormalized data?
Regards
SN

[franzbischoff]

Hello, Please try my new package (that will be a dependency of tsmp): install this way the development version: remotes::install_github("matrix-profile-foundation/matrixprofiler@release/Release-1") best regards.

[vanbenschoten]

@snvv I'd also recommend comparing against our Python implementation "matrixprofile". That could serve as a tie-breaker in this case :)