ai.libs.jaicore.ml.tsc.distances

Class TimeWarpEditDistance

java.lang.Object

ai.libs.jaicore.ml.tsc.distances.TimeWarpEditDistance

All Implemented Interfaces:

ITimeSeriesDistance, ITimeSeriesDistanceWithTimestamps

public class TimeWarpEditDistance
extends java.lang.Object
implements ITimeSeriesDistanceWithTimestamps

Time Warp Edit Distance as published in "Time Warp Edit Distance with Stiffness Adjustment for Time Series Matching" by Pierre-Francois Marteau (2009). The similarity between two time series is measured as the minimum cost sequence of edit operations needed to transform one time series into another.

Constructor Summary

Constructors

Constructor and Description
TimeWarpEditDistance(double lambda, double nu) Constructor.
TimeWarpEditDistance(double lambda, double nu, IScalarDistance d) Constructor.

Method Summary

Modifier and Type	Method and Description
double	distance(double[] a, double[] tA, double[] b, double[] tB) Calculates the distance between two time series.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface ai.libs.jaicore.ml.tsc.distances.ITimeSeriesDistanceWithTimestamps

distance

Constructor Detail

TimeWarpEditDistance

public TimeWarpEditDistance(double lambda,
                            double nu,
                            IScalarDistance d)

Constructor.

Parameters:

lambda - Additional cost parameter for deletion.

nu - Stiffness parameter. Used to parametrize the influence of the time stamp distance.

d - Distance mesaure used for point distance calculation.

TimeWarpEditDistance

public TimeWarpEditDistance(double lambda,
                            double nu)

Constructor. Creates a TimeWarpEditDistance with squared distance as point distance.

Parameters:

lambda - Additional cost parameter for deletion.

nu - Stiffness parameter.

Method Detail

distance

public double distance(double[] a,
                       double[] tA,
                       double[] b,
                       double[] tB)

Description copied from interface: ITimeSeriesDistanceWithTimestamps

Calculates the distance between two time series.

Specified by:

distance in interface ITimeSeriesDistanceWithTimestamps

Parameters:

a - First time series.

tA - Timestamps for the first time series.

b - Second time series.

tB - Timestamps for the second times series.

Returns:

Distance between the first and second time series.