DTW

scalation.modeling.forecasting.DTW
class DTW(y: VectorD, q: Int)(ty: VectorT)

The DTW class is used for aligning two time series.

Value parameters

q

use the L_q norm (defaults to 2 (Euclidean))

ty

the corresponding date-time vector

y

the first time series vector

Attributes

Graph
Supertypes
class Object
trait Matchable
class Any

Members list

Value members

Concrete methods

def align(z: VectorD)(tz: VectorT): (VectorD, VectorD)

Align the two time series.

Align the two time series.

Value parameters

tz

the corresponding date-time vector

z

the second time series vector

Attributes

def d(a: Double, b: Double): Double

Return the difference between a and b.

Return the difference between a and b.

Value parameters

a

the first value

b

the second value

Attributes

def dtw_score(z: VectorD, w: Int): Double

Return an alignment score for the two time series y and z. Translated from @see rtavenar.github.io/blog/dtw.html Note that the 3 elements in VectorD cannot all be infinity if we have (i > 0 or j > 0)

Return an alignment score for the two time series y and z. Translated from @see rtavenar.github.io/blog/dtw.html Note that the 3 elements in VectorD cannot all be infinity if we have (i > 0 or j > 0)

Value parameters

w

the size of the locally window (to limit the index span between index i and j)

z

the second time series vector

Attributes