Matrix Profile XIII: Time Series Snippets: A New Primitive for Time Series Data Mining

Perhaps the most basic query made by a data analyst confronting a new data source is "Show me some representative/typical data." Answering this question is trivial in many domains, but surprisingly, it is very difficult in large time series datasets. The major difficulty is not time or space complexity, but defining what it means to be representative data in this domain. In this work, we show that the obvious candidate definitions: motifs, shapelets, cluster centers, random samples etc., are all poor choices. Thus motivated, we introduce time series snippets, a novel representation of typical time series subsequences. Beyond their utility for visualizing and summarizing massive time series collections, we show that time series snippets have utility for high-level comparison of large time series collections.