DTWDist_fast: (Fast) Dynamic Time Warping Dissimilarity

In Jakob-Bach/FastTSDistances: Fast dissimilarity computations for time series

Description

Fast version of univariate dynamic time warping (unconstrained, symmetric1 step pattern) which uses a cyclic access strategy with a smaller cost matrix; inspired by the C implementation of dynamic time warping in dtwclust::dtw_basic(), but cuts even more overhead.

Usage

DTWDist_fast(x, y, cid = FALSE, cortK = -1, normalize = FALSE)

Arguments

x	1st numeric vector/time series.
y	2nd numeric vector/time series.
cid	Should the distance be made "complexity-invariant" (l2CompCorFactor_fast)?
cortK	Should the temporal behavior (correlation) of the time series' diff vectors be considered (cortFactor_fast)? A factor smaller than 0 means no, higher factors will be used as parameter k in the temporal correlation scaling function.
normalize	Divide by the length of the longer time series (= minimum amount of assignment steps) to account for series of different lengths in your dataset.

Details

Be aware that it is not really a distance in the strict sense, as DTW violates the triangle inequality.

Value

The distance as double.

References

Berndt, D. J. & Clifford, J. (1994). Using dynamic time warping to find patterns in time series. In Proceedings of the 3rd international conference on knowledge discovery and data mining (pp. 359–370). AAAI Press.

See Also

Other DTW functions: DTWDistMult_fast, DTWDistSakoeChibaMult_fast, DTWDistSakoeChiba_fast

Jakob-Bach/FastTSDistances documentation built on May 13, 2019, 1:15 p.m.