Package ai.libs.jaicore.ml.classification.singlelabel.timeseries.shapelets.search

Class EarlyAbandonMinimumDistanceSearchStrategy

  • public class EarlyAbandonMinimumDistanceSearchStrategy
extends AMinimumDistanceSearchStrategy
    Class implementing a search strategy used for finding the minimum distance of a Shapelet object to a time series. The approach uses early abandoning as described in algorithm 2 in the paper 'Jason Lines, Luke M. Davis, Jon Hills, and Anthony Bagnall. 2012. A shapelet transform for time series classification. In Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining (KDD '12). ACM, New York, NY, USA, 289-297.'.

    • Constructor Detail

    • Method Detail

      • findMinimumDistance

        public double findMinimumDistance​(Shapelet shapelet,
                                  double[] timeSeries)
        Optimized function returning the minimum distance among all subsequences of the given timeSeries to the shapelet's data. This function implements the algorithm 2 mentioned in the original paper. It performs the similarity search with online normalization and early abandon.
        Specified by:
        findMinimumDistance in class AMinimumDistanceSearchStrategy
        Parameters:
        shapelet - The shapelet to be compared to all subsequences
        timeSeries - The time series which subsequences are compared to the shapelet's data
        Returns:
        Return the minimum distance among all subsequences