How to use the UCR Matrix Profile Algorithm for Motif Mining in R I am seeking to understand how to use the mstomp_par (Multivariate STOMP algorithm Parallel version) to detect multivariate time series motifs. As of now, I see that this algorithm provides the matrix profile and profile index. I would like to know how to use this information to isolate motifs from data.
 A: In R, this system of tools is provided by the package, tsmp, which is both on CRAN and github, currently at version 3.4.9. If you look at the help for mstomp_par, you'll see that you provide a column in a matrix,data_frame, or list for each of your multivariates, set your conditions for window size & etc, though after window size defaults are pretty good. 
It returns a MultiMatrixProfile object, a list with the matrix profile mp, profile index pi left and right matrix profile lmp, rmp and profile index lpi, rpi, window size w, number of dimensions n_dim, exclusion zone ez, must dimensions must and excluded dimensions exc.
The matrix profile and profile index are keys to many kingdoms. With the matrix profile you can go on to find motifs, chains, discords. For 'how to do the coding questions in R' (or python or C or ObjectC), I would go over to Stackoverflow.com,
where, unfortunately there isn't that much on matrix profile in R as yet. But starting a discussion starts a discussion, and people will answer.   
