Trainable Time Warping: Aligning Time-Series in the Continuous-Time Domain
This addresses the need for a fast, high-quality multisequence alignment algorithm for researchers and practitioners dealing with large-scale time-series data, representing a novel method rather than an incremental improvement.
The paper tackles the problem of aligning multiple time-series efficiently by introducing Trainable Time Warping (TTW), which achieves linear computational complexity in both the number and length of time-series and outperforms Generalized Time Warping (GTW) on 67.1% of datasets for averaging tasks and 61.2% for classification tasks.
DTW calculates the similarity or alignment between two signals, subject to temporal warping. However, its computational complexity grows exponentially with the number of time-series. Although there have been algorithms developed that are linear in the number of time-series, they are generally quadratic in time-series length. The exception is generalized time warping (GTW), which has linear computational cost. Yet, it can only identify simple time warping functions. There is a need for a new fast, high-quality multisequence alignment algorithm. We introduce trainable time warping (TTW), whose complexity is linear in both the number and the length of time-series. TTW performs alignment in the continuous-time domain using a sinc convolutional kernel and a gradient-based optimization technique. We compare TTW and GTW on 85 UCR datasets in time-series averaging and classification. TTW outperforms GTW on 67.1% of the datasets for the averaging tasks, and 61.2% of the datasets for the classification tasks.