Crate dtw 

Implementation of dynamic time warping (DTW) algorithm and approximations.

The DTW algorithm [1] finds the optimal alignment between two time series.

This crate provides a DTW implementations as well as a FastDTW [2] implementation. FastDTW is linear time and space approximation of DTW.

References

[1] Kruskal, JB & Liberman, Mark. (1983). The symmetric time-warping problem: From continuous to discrete. Time Warps, String Edits, and Macromolecules: The Theory and Practice of Sequence Comparison.

[2] Salvador, Stan & Chan, Philip. (2004). Toward Accurate Dynamic Time Warping in Linear Time and Space. Intelligent Data Analysis. 11. 70-80.

Modules

dist

Usual distance functions.

Structs

Constraint

A constraint for the DTW search space, where each row is constrained to a single contiguous segment.

RowConstraint

A constraint constraining a row to a closed line segment.

Traits

Average

Types that can be averaged.

SumContainer

Provide an associated type to store sum.

Functions

constrained_dtw

Compute the warp path of two given time series using DTW [1] with a constrained search space.

constrained_dtw_with_cmp

Compute the warp path of two given time series using DTW [1] with a constrained search space.

dtw

Compute the warp path of two given time series using DTW [1].

dtw_with_cmp

Compute the warp path of two given time series using DTW [1].

fast_dtw

Compute the warp path of two given time series using FastDTW [1].

fast_dtw_with_cmp

Compute the warp path of two given time series using FastDTW [1].