Support for multivariate time series

[ortegatron]

Hi, I was looking for motif discovery on timeseries and find this one. Seems very good, but my data has several variables, it's a multivariate time serie. This library doesn't support that, right? Would it be easy to change the code to support it, or it is just non viable?

Thanks, MArcelo

[vanbenschoten]

Hi Marcelo, It's a non-trivial addition, but do-able (we just haven't had the bandwidth to incorporate it yet). We'll hopefully include it in the near future, though you're always welcome to submit your own PR :) In the meantime, the other two libraries supported by the Matrix Profile Foundation (go-matrixprofile (Golang) and tsmp (R)) have multivariate capabilities.

[ortegatron]

Sweet, gonna have a look at that ones. Thank you for the quick reply!

[aouyang1]

@ortegatron You can also take a look at this paper for multivariate mp. https://www.cs.ucr.edu/~eamonn/Motif_Discovery_ICDM.pdf.

[Sandy4321]

great code thanks may you still clarify : will it work for multivariate time series as written in Matrix Profile VI: Meaningful Multidimensional Motif Discovery https://www.cs.ucr.edu/~eamonn/Motif_Discovery_ICDM.pdf. 1 where all values are continues values 2 or even will it work for multivariate time series where values are mixture of continues and categorical values for example 2 dimensions have continues values and 3 dimensions are categorical values

color        weight     gender  height  age  

1 black 56 m 160 34 2 white 77 f 170 54 3 yellow 87 m 167 43 4 white 55 m 198 72 5 white 88 f 176 32

[vanbenschoten]

@Sandy4321 unfortunately this library is no longer maintained. I'd recommend you take a look at the Matrix Profile Foundation's new library here. It's substantially more performant than matrixprofile-ts, and includes a number of additional features (pan-Matrix Profile, Snippets support, etc).

With respect to your original question, the matrixprofile library has an open issue to add multi-dimensional support in the near future. Feel free to add to the discussion there, as well as submit a PR :)