Statistica Dynamic Time Warping

The Dynamic Time Warping (DTW) node was introduced in Statistica 13.4. 

Dynamic Time Warping (DTW) is a popular technique for optimally aligning two or multiple time-dependent data sources.  The technique was originally used to compare different speech patterns in automatic speech recognition.

It has statistical applications as a data-preprocessing technique e.g. MSPC analyses require all batch trajectory data to have an equal number of time points. DTW can be used to align such batch data.

The node allows the user to apply one of the following distance functions as a cost function.

• Euclidean distance
• Squared Euclidean distance
• Manhattan distance
• Chebyshev distance
• Power distance with user defined p and q where p, q ? {1,2,3,4}

References

Chapter 4 of Information Retrieval for Music and Motion by Meinard Müller is a good overview of this topic. This was published in 2007. 

Athanassios Kassidas, John F. MacGregor, Paul A. Taylor, Synchronization of batch trajectories using dynamic time warping. American Institute of Chemical Engineers (AIChE), Volume44, Issue4, April 1998, Pages 864-875

F. Itakura, Minimum Prediction Residual Principle Applied to Speech Recognition. IEEE Transactions on Acoustics, Speech, and Signal Processing, Volume: 23 , Issue: 1 , Feb 1975, Pages 154-158 

Abdullah Mueen, Eamonn J. Keogh: Extracting Optimal Performance from Dynamic Time Warping. KDD 2016: 2129-2130