Continuous, Online Anomaly Region Detection and Tracking in Networks

  • Shuiyuan Xie
  • Xiuli Ma
  • Shiwei Tang
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7901)


In many real networks, the detection and tracking of unusual phenomena, such as the diffusion of contamination and the spreading of disease, is one of the key feature users are great interested in, which is called anomaly with technical terms. In this paper, we present a framework to detect and track anomaly region continuously. First, we build a state transition graph to summarize network’s operating regularity, that is, network stays in a state for a period of time and alternates among states over and over again, which exists in many real networks. Second, we employ the state transition graph to predict network’s next state. While comparing expected state and current state, we present suspicious region and its anomaly probability. We evaluate our approach on a real water distribution network from the Battle of the Water Sensor Network (BWSN). Experiments show that our approach is effective, efficient and scalable to detect and track anomaly region.


Anomaly region detection and tracking state transition graph 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Sakaki, T., Okazaki, M., Matsuo, Y.: Earthquake Shakes Twitter Users: Real-time Event Detection by Social Sensors. In: Proc. of the 19th International Conference on World Wide Web, pp. 851–860 (2010)Google Scholar
  2. 2.
    Liu, W., Zheng, Y., Chawla, S., Yuan, J., Xie, X.: Discovering Spatio-Temporal Causal Interactions in Traffic Data Streams. In: Procedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 1010–1018 (2011)Google Scholar
  3. 3.
    Franke, C., Gertz, M.: Outlier Region Detection and Exploration in Sensor Networks. In: Proc. of the 2009 ACM SIGMOD International Conference on Management of Data, pp. 1075–1078 (2009)Google Scholar
  4. 4.
    Xue, W., Luo, Q., Chen, L., Liu, Y.: Contour Map Matching for Event Detection in Sensor Networks. In: Proc. of the 2006 ACM SIGMOD International Conference on Management of Data, pp. 145–156 (2006)Google Scholar
  5. 5.
    Wang, P., Wang, H., Wang, W.: Finding Semantics in Time Series. In: Proc. of the 2011 ACM SIGMOD International Conference on Management of Data, pp. 385–396 (2011)Google Scholar
  6. 6.
    Rossman, L.A.: EPANET2 user’s manual: National Risk Management Research Laboratory, U.S. Environmental Protection Agency (2000)Google Scholar
  7. 7.
    Xiao, H., Ma, X., Tang, S., Tian, C.: Continuous Summarization of Co-Evolving Data in Large Water Distribution Network. In: Chen, L., Tang, C., Yang, J., Gao, Y. (eds.) WAIM 2010. LNCS, vol. 6184, pp. 62–73. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  8. 8.
    Chi, Y., Yu, P.S., Wang, H., Muntz, R.R.: Loadstar: A Load Shedding Scheme for Classifying Data Streams. In: Proc. of the 31st International Conference on Very Large Data Bases, pp. 1302–1305 (2005)Google Scholar
  9. 9.
    Ostfeld, A., Uber, J.G., Salomons, E.: Battle of water sensor networks: A design challenge for engineers and algorithms. In: WDSA (2006)Google Scholar
  10. 10.
    Hart, D.B., Klise, K.A., McKenna, S.A., Wilson, M.P.: CANARY User’s Manual Version 4.1. Sandia National Laboratories. U.S. Environmental Protection Agency (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Shuiyuan Xie
    • 1
    • 2
  • Xiuli Ma
    • 1
    • 2
  • Shiwei Tang
    • 1
    • 2
  1. 1.Key Laboratory of Machine Perception, Ministry of EducationPeking UniversityChina
  2. 2.School of Electronics Engineering and Computer SciencePeking UniversityBeijingChina

Personalised recommendations