Skip to Main Content
Similarity-based time-series retrieval has been a subject of long-term study due to its wide usage in many applications, such as financial data analysis, weather data forecasting, and multimedia data retrieval. Its original task was to find those time series similar to a pattern (query) time-series data, where both the pattern and data time series are static. Recently, with an increasing demand on stream data management, similarity-based stream time-series retrieval has raised new research issues due to its unique requirements during the stream processing, such as one-pass search and fast response. In this paper, we address the problem of matching both static and dynamic patterns over stream time-series data. We will develop a novel multiscale representation, called multiscale segment mean, for stream time-series data, which can be incrementally computed and thus perfectly adapted to the stream characteristics. Most importantly, we propose a novel multistep filtering mechanism, step by step, over the multiscale representation. Analysis indicates that the mechanism can greatly prune the search space and thus offer fast response. Furthermore, batch processing optimization and the dynamic case where patterns are also from stream time series are discussed. Extensive experiments show the multiscale representation together with the multistep filtering scheme can efficiently filter out false candidates and detect patterns, compared to the multiscale wavelet.