Experiencing the Shotgun Distance for Time Series Analysis

Similarity search is a core functionality in many data mining algorithms. Over the past decade algorithms were designed to mostly work with human assistance to extract characteristic, aligned patterns of equal length and scaling. We propose the shotgun distance similarity measure that extracts, scales, and aligns segments from a query to a sample time series. This greatly simplifies the time series analysis task of those time series produced by sensors. We show the applicability of our shotgun distance in the context of hierarchical clustering of heraldic shields, and human motion detection. A time series is segmented using varying lengths as part of our shotgun ensemble classifier. This classifier improves the best published accuracies on case studies in the context of bioacoustics, human motion detection, spectrographs or personalized medicine. Finally, it performs better than state of the art on the official UCR classification benchmarks.