Time Series Motif Discovery Based on Subsequence Join under Dynamic Time Warping
Luong Van Do, Duong Tuan Anh · 2017
Time series motifs are repeated similar subsequences in a longer time series. Finding motif in time series is useful in many practical applications. In this paper, we propose a novel approach for time series motif discovery. This approach first uses subsequence join to obtain the similarity relationships among subsequences in a time series. Our subsequence join method is based on segmentation and matching under Dynamic Time Warping. Then, the motif discovery problem can be converted to a maximum clique problem solvable by some efficient graph-theoretic algorithm. Experiments demonstrate the effectiveness and efficiency of the proposed approach to discovery motifs in real-world time series data.