wannesm/dtaidistance v2.0.0

Wannes Meert, Kilian Hendrickx, Toon Van CraenendonckZenodo 2020 · DOI

Abstract

New in v2: Numpy is now an optional dependency, also to compile the C library (only Cython is required). Small optimizations throughout the C code to improve speed. The consistent use of <code>size_t</code> instead of <code>int</code> allows for larger data structures on 64 bit machines and be more compatible with Numpy. The parallelization is now implemented directly in C (included if OpenMP is installed). The <code>max_dist</code> argument turned out to be similar to Silva and Batista's work on PrunedDTW [7]. The toolbox now implements a version that is equal to PrunedDTW since it prunes more partial distances. Additionally, a <code>use_pruning</code> argument is added to automatically set <code>max_dist</code> to the Euclidean distance, as suggested by Silva and Batista, to speed up the computation (a new method <code>ub_euclidean</code> is available). Support in the C library for multi-dimensional sequences in the <code>dtaidistance.dtw_ndim</code> package.
BibTeX
@inproceedings{meert2020wannesm,
  author = {Wannes Meert and Kilian Hendrickx and Toon Van Craenendonck},
  title = {{wannesm/dtaidistance v2.0.0}},
  booktitle = {Zenodo},
  year = {2020},
  doi = {10.5281/zenodo.3981067},
}

Loading related content...

Figure