Efficient Detection of Discords for Time Series Stream
Time discord detection is an important problem in a great variety of applications. In this paper, we consider the problem of discord detection for time series stream, where time discords are detected from local segments of flowing time series stream. The existing detections, which aim to detect the global discords from time series database, fail to detect such local discords. Two online detection algorithms are presented for our problem. The first algorithm extends the existing algorithm HOT SAX to detect such time discords. However, this algorithm is not efficient enough since it needs to search the entire time subsequences of local segment. Then, in the second algorithm, we limit the search space to further enhance the detection efficiency. The proposed algorithms are experimentally evaluated using real and synthesized datasets.
KeywordsTime discord Online detection Time series stream
Unable to display preview. Download preview PDF.
- 1.Fu, A., Keogh, E., Lau, L.Y.H., Ratanamahatana, C.A.: Scaling and time warping in time series querying. In: Proc. VLDB, pp. 649–660 (2005)Google Scholar
- 2.Keogh, E., Lin, J., Fu, A.: HOT SAX: efficiently finding the most unusual time series subsequence. In: Proc. ICDM, pp. 226–233 (2005)Google Scholar
- 3.Bu, Y., Leung, T.-W., Fu, A., Keogh, E., Pei, J., Meshkin, S.: WAT: Finding Top-K Discords in Time Series Database. In: Proc. SDM, pp. 449–454 (2007)Google Scholar
- 4.Yankov, D., Keogh, E., Rebbapragad, U.: Disk Aware Discord Discovery: Finding Unusual Time Series in Terabyte Sized Datasets. In: Proc. ICDM 2007 (2007)Google Scholar
- 5.Chan, K.-P., Fu, A.: Efficient time series matching by wavelets. In: Proc. ICDE, pp. 126–133 (1999)Google Scholar
- 8.Wang, H., Yin, J., Pei, J., Yu, P.S., Yu, J.X.: Suppressing model over-fitting in mining concept-drifting data streams. In: Proc. KDD, pp. 736–741 (2006)Google Scholar
- 9.Qin, S., Qian, W., Zhou, A.: Approximately processing multi-granularity aggregate queries over data streams. In: Proc. ICDE (2006)Google Scholar