Back to models
Transformer (Pairwise Panel)

SoftDtwDistTslearn

API referenceSource
MultivariateUnequal lengthMissing values

Soft dynamic time warping distance, from tslearn.

Quickstart

python
from sktime.dists_kernels.dtw import SoftDtwDistTslearn

estimator = SoftDtwDistTslearn(normalized=False, gamma=1.0)

Parameters(2)

normalizedbool, default = False

Whether the DTW distance should be normalized. If False, interfaces tslearn.metrics.cdist_soft_dtw. If True, interfaces tslearn.metrics.cdist_soft_dtw_normalized.

gammafloat (default 1.)
Gamma parameter for Soft-DTW

References

  1. [1 ] M. Cuturi, M. Blondel “Soft-DTW: a Differentiable Loss Function for Time-Series,” ICML 2017.