This paper introduces a modified version of Dynamic Time Warping (DTW), called Piecewise Dynamic Time Warping (PDTW), which allows for faster computation while maintaining accuracy. The authors demonstrate that Euclidean distance, commonly used for time series comparison, is highly sensitive to small distortions in the time axis and can produce unintuitive results. DTW, which allows for more robust distance calculations by aligning sequences in time, is computationally expensive. PDTW improves upon DTW by using a higher-level abstraction of the data, specifically a Piecewise Aggregate Approximation (PAA), which reduces the dimensionality of the time series while preserving key characteristics. This approach enables PDTW to outperform DTW by one to two orders of magnitude in speed without sacrificing accuracy. The paper also presents experimental results showing that PDTW performs well on real-world datasets, including the Australian Sign Language (ASL) dataset and the Cylinder-Bell-Funnel (CBF) synthetic dataset. The results indicate that PDTW achieves high accuracy and significantly faster computation compared to DTW and Euclidean distance. The authors conclude that PDTW is a valuable tool for time series data mining applications where speed and accuracy are both important.This paper introduces a modified version of Dynamic Time Warping (DTW), called Piecewise Dynamic Time Warping (PDTW), which allows for faster computation while maintaining accuracy. The authors demonstrate that Euclidean distance, commonly used for time series comparison, is highly sensitive to small distortions in the time axis and can produce unintuitive results. DTW, which allows for more robust distance calculations by aligning sequences in time, is computationally expensive. PDTW improves upon DTW by using a higher-level abstraction of the data, specifically a Piecewise Aggregate Approximation (PAA), which reduces the dimensionality of the time series while preserving key characteristics. This approach enables PDTW to outperform DTW by one to two orders of magnitude in speed without sacrificing accuracy. The paper also presents experimental results showing that PDTW performs well on real-world datasets, including the Australian Sign Language (ASL) dataset and the Cylinder-Bell-Funnel (CBF) synthetic dataset. The results indicate that PDTW achieves high accuracy and significantly faster computation compared to DTW and Euclidean distance. The authors conclude that PDTW is a valuable tool for time series data mining applications where speed and accuracy are both important.