Package ai.libs.jaicore.basic.metric

Class WeightedDynamicTimeWarping

  • All Implemented Interfaces:
    org.api4.java.common.math.IMetric<double[]>, IDistanceMetric
    public class WeightedDynamicTimeWarping
extends java.lang.Object
implements IDistanceMetric
    Implementation of the Dynamic Time Warping (DTW) measure as published in "Weighted dynamic time warping for time series classification" by Young-Seon Jeong, Myong K. Jeong and Olufemi A. Omitaomu. DTW does not account for the relative importance regarding the phase difference between a reference point and a testing point. This may lead to misclassification especially in applications where the shape similarity between two sequences is a major consideration for an accurate recognition. Therefore, [the authors] propose a novel distance measure, called a weighted DTW (WDTW), which is a penalty-based DTW. [Their] approach penalizes points with higher phase difference between a reference point and a testing point in order to prevent minimum distance distortion caused by outliers.

    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      protected double[] calculateWeights​(int length)
      Calculates the weight vector via the Modified logistic weight function (see paper 4.2).
      double distance​(double[] a, double[] b)
      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

    • Constructor Detail

      • WeightedDynamicTimeWarping

        public WeightedDynamicTimeWarping​(double g,
                                  double maximumWeight,
                                  IScalarDistance d)
        Constructor.
        Parameters:
        g - Controls the penelization in weights for points with larger phase difference.
        maximumWeight - The desired upper bound for the weight parameter that is used to penalize points with higher phase difference.

    • Method Detail

      • distance

        public double distance​(double[] a,
                       double[] b)
        Description copied from interface: IDistanceMetric
        Calculates the distance between two time series.
        Specified by:
        distance in interface IDistanceMetric
        Parameters:
        a - First time series.
        b - Second time series.
        Returns:
        Distance between the first and second time series.

      • calculateWeights

        protected double[] calculateWeights​(int length)
        Calculates the weight vector via the Modified logistic weight function (see paper 4.2). Uses memoization to avoid multiple calculations for the same length.
        Parameters:
        length - Length of the time series, i.e. length of the weight vector. Is guaranteed to be greater 0 within this class.
        Returns:
        Resulting weight vector.