Presenting 'matrixprofiler' a fast Matrix Profile implementation in R

Posts

It took some time, as you can see in the previous post, but the matrixprofiler package is done!

What does this mean?

The UCR Matrix Pofile is growing, undoubtedly, and the tsmp package is getting almost 700 downloads per month. We decided that we needed to separate the core from the practical usage of Matrix Profile.

So the matrixprofiler package has been born and is focused on having the low-level code (C/C++), for speed and robustness. On top of that, the current tsmp package will focus on implementing the algorithms that use the matrix profile for actual data science, in a high-level code (R language), easier to understand to the actual user/scientist, using as leverage the speed of matrixprofiler. You can see the benchmarks here.

This new package is not intended to be used directly (but as a dependency of tsmp), but it can be. Just be careful to provide the required data types, and everything will work fine.

In addition to the core algorithms: mass, stamp, stomp, etc; this package offers fast and accurate implementation of several math and window functions like movmean, movmax, zero_crossing, and will in the future provide more basic tools to be used by the high-level algorithms.

Now starts the tsmp phase of getting rid of the core algorithms and leverage its new dependency: matrixprofiler.