Making the dynamic time warping distance warping-invariant
作者：
Highlights：
• The dtw-distance is not warping-invariant.
• Lack of warping-invariance contributes to inconsistencies and peculiarities.
• Converts dtw-distance into a warping-invariant semi-metric.
• Proposed semi-metric is more efficient than dtw-distance.
• Error-rates of nn-classifiers in semi-metric and dtw-space are practically equivalent.
摘要
•The dtw-distance is not warping-invariant.•Lack of warping-invariance contributes to inconsistencies and peculiarities.•Converts dtw-distance into a warping-invariant semi-metric.•Proposed semi-metric is more efficient than dtw-distance.•Error-rates of nn-classifiers in semi-metric and dtw-space are practically equivalent.
论文关键词：Time series,Dynamic time warping,Semi-metric,Nearest-neighbor rule
论文评审过程：Received 17 October 2018, Revised 2 March 2019, Accepted 8 May 2019, Available online 11 May 2019, Version of Record 20 May 2019.
论文官网地址：https://doi.org/10.1016/j.patcog.2019.05.016
