Clustering method for time-series images using quantum-inspired digital annealer technology

Abstract

Time-series clustering is a powerful data mining technique for time-series data in the absence of prior knowledge of the clusters. Here we propose a time-series clustering method that leverages an annealing machine, which accurately solves combinatorial optimization problems. The proposed method facilitates an even classification of time-series data into closely located clusters while maintaining robustness against outliers. We compared the proposed method with an existing standard method for clustering an online distributed dataset and found that both methods yielded comparable results. Furthermore, the proposed method was applied to a flow measurement image dataset containing noticeable noise with a signal-to-noise ratio of approximately unity. Despite a small signal variation of approximately 2%, the proposed method effectively classified the data without any overlaps among the clusters. In contrast, the clustering results of the existing methods exhibited overlapping clusters. These results indicate the effectiveness of the proposed method.