Package adams.data.filter

Class TimeseriesSAX

  • All Implemented Interfaces:
    ClassCrossReference, CleanUpHandler, CrossReference, Destroyable, GlobalInfoSupporter, LoggingLevelHandler, LoggingSupporter, OptionHandler, ShallowCopySupporter<Filter>, SizeOfHandler, TechnicalInformationHandler, Filter<Timeseries>, Serializable, Comparable
    public class TimeseriesSAX
extends AbstractSAX<Timeseries>
implements ClassCrossReference
    Performs Symbolic Aggregate approXimation (SAX).
    The data must be normalized using adams.data.filter.RowNorm beforehand.
    For more information see:

    Chiu, B., Keogh, E., Lonardi, S. (2003). Probabilistic Discovery of Time Series Motifs.

    See also:
    adams.data.filter.RowNorm 

     @proceedings{Chiu2003,
    author = {Chiu, B. and Keogh, E. and Lonardi, S.},
    booktitle = {9th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
    pages = {493-498},
    title = {Probabilistic Discovery of Time Series Motifs},
    year = {2003},
    location = {Washington, DC, USA},
    PDF = {http://www.cs.ucr.edu/\~eamonn/SIGKDD_Motif.pdf}
 }
 


    -logging-level <OFF|SEVERE|WARNING|INFO|CONFIG|FINE|FINER|FINEST> (property: loggingLevel)
    The logging level for outputting errors and debugging output.
    default: WARNING
 
    -no-id-update <boolean> (property: dontUpdateID)
    If enabled, suppresses updating the ID of adams.data.id.IDHandler data containers.
    default: false
 
    -num-windows <int> (property: numWindows)
    The number of windows to use for Piecewise Aggregate Approximation (PAA).
    default: 10
    minimum: 1
 
    -num-bins <int> (property: numBins)
    The number of bins to use for the Gaussian.
    default: 10
    minimum: 1
 
    -output-labels <boolean> (property: outputLabels)
    If enabled, labels are output instead of distances.
    default: true
    Version:
    $Revision: 1286 $
    Author:
    fracpete (fracpete at waikato dot ac dot nz)
    See Also:
    Serialized Form

    • Constructor Detail

      • TimeseriesSAX

        public TimeseriesSAX()

    • Method Detail

      • getMeanDeltaX

        protected double getMeanDeltaX​(Timeseries data)
        Computes the mean difference between data points on the X axis.
        Specified by:
        getMeanDeltaX in class AbstractSAX<Timeseries>
        Parameters:
        data - the data to use for the calculation
        Returns:
        the mean

      • getX

        protected double getX​(DataPoint point)
        Obtains the X value from the given data point.
        Specified by:
        getX in class AbstractSAX<Timeseries>
        Parameters:
        point - the data point to extract the X value from
        Returns:
        the X value

      • getY

        protected double getY​(DataPoint point)
        Obtains the Y value from the given data point.
        Specified by:
        getY in class AbstractSAX<Timeseries>
        Parameters:
        point - the data point to extract the Y value from
        Returns:
        the Y value

      • newDataPoint

        protected DataPoint newDataPoint​(double x,
                                 double y)
        Creates a new data point from the X and Y values.
        Specified by:
        newDataPoint in class AbstractSAX<Timeseries>
        Parameters:
        x - the raw X value
        y - the raw Y value
        Returns:
        the data point