{"title":"Evidence Theory Enabled Quickest Change Detection Using Big Time-Series Data from Internet of Things","authors":"Hossein Jafari, Xiangfang Li, Lijun Qian, Alexander Aved, Timothy Kroecker","volume":126,"journal":"International Journal of Computer and Information Engineering","pagesStart":710,"pagesEnd":717,"ISSN":"1307-6892","URL":"https:\/\/publications.waset.org\/pdf\/10007423","abstract":"Traditionally in sensor networks and recently in the
\r\nInternet of Things, numerous heterogeneous sensors are deployed
\r\nin distributed manner to monitor a phenomenon that often can be
\r\nmodel by an underlying stochastic process. The big time-series
\r\ndata collected by the sensors must be analyzed to detect change
\r\nin the stochastic process as quickly as possible with tolerable
\r\nfalse alarm rate. However, sensors may have different accuracy
\r\nand sensitivity range, and they decay along time. As a result,
\r\nthe big time-series data collected by the sensors will contain
\r\nuncertainties and sometimes they are conflicting. In this study, we
\r\npresent a framework to take advantage of Evidence Theory (a.k.a.
\r\nDempster-Shafer and Dezert-Smarandache Theories) capabilities of
\r\nrepresenting and managing uncertainty and conflict to fast change
\r\ndetection and effectively deal with complementary hypotheses.
\r\nSpecifically, Kullback-Leibler divergence is used as the similarity
\r\nmetric to calculate the distances between the estimated current
\r\ndistribution with the pre- and post-change distributions. Then mass
\r\nfunctions are calculated and related combination rules are applied to
\r\ncombine the mass values among all sensors. Furthermore, we applied
\r\nthe method to estimate the minimum number of sensors needed to
\r\ncombine, so computational efficiency could be improved. Cumulative
\r\nsum test is then applied on the ratio of pignistic probability to detect
\r\nand declare the change for decision making purpose. Simulation
\r\nresults using both synthetic data and real data from experimental
\r\nsetup demonstrate the effectiveness of the presented schemes.","references":"[1] Basseville, M. and Nikiforov, I. V. Detection of Abrupt Change Theory\r\nand Application, Englewood Cliffs, NJ: Prentice Hall, 1993.\r\n[2] Poor, H. V. and Hadjiliadis, O. Quickest Detection, New York: Cambridge\r\nUniversity Press, 2008.\r\n[3] Page, E. S. Continuous inspection schemes, Biometrika, 1954.\r\n[4] Shiryaev, A. N. On Optimum Methods in Quickest Detection Problems,\r\nTheory of Probability and Its Applications 8: 22\u201346, 1963.\r\n[5] A. G. Tartakovsky, B. Rozovskii, R. Blazek, and H. Kim, A Novel\r\nApproach to Detection of Intrusions in Computer Networks via Adaptive\r\nSequential and Batch-Sequential Change-point Detection Methods, IEEE\r\nTrans. Sig. Proc., vol. 54, no. 9, pp. 3372\u20133382, Sep. 2006.\r\n[6] J. S. Baras, A. Cardenas, and V. Ramezani, Distributed Change Detection\r\nfor Worms, DDOS and Other Network Attacks, Proc. American Cont.\r\nConf. (ACC), Boston, MA, pp. 1008\u20131013, 2004.\r\n[7] Husheng Li, Chengzhi Li and Huaiyu Dai, Quickest spectrum sensing in\r\ncognitive radio, Information Sciences and Systems. CISS. 42nd Annual\r\nConference on, Princeton, NJ, pp. 203\u2013208, 2008.\r\n[8] B. Khaleghi, A. Khamis, F. O. Karray, and S. N. Razavi, Multisensor\r\ndata fusion: A review of the state-of-the-art, Information Fusion, vol. 14,\r\nno. 1, pp. 28\u201344, 2013.\r\n[9] Tartakovsky, A. G. and Veeravalli, V. V. Asymptotically optimal quickest\r\nchange detection in distributed sensor systems, Sequential Anal. 27 pp.\r\n441\u2013475, 2008.\r\n[10] Mei, Y. Efficient scalable schemes for monitoring a large number of\r\ndata streams, Biometrika 97 pp. 419\u2013433, 2010.\r\n[11] Yao Xie and David Siegmund. Sequential multi-sensor change-point\r\ndetection, in The Annals of Statistics, Vol. 41, No. 2, pp. 670\u2013692, 2013.\r\n[12] G. Shafer, Perspectives on the theory and practice of belief functions,\r\nInternational Journal of Approximate Reasoning, vol. 4, no. 5, pp.\r\n323\u2013362, 1990.\r\n[13] F. Smarandache and J. Dezert, Advances and Applications of DSmT for\r\nInformation Fusion, in American Research Press, Rehoboth, vol. 1-2,\r\n2006.\r\n[14] R. R. Yager and L. Liu, Classic Works of the Dempster-Shafer Theory\r\nof Belief Functions, in American Research Press, Rehoboth, vol. 2, ch. 1,\r\npp. 3\u201368, 2006.\r\n[15] P. Smets, Constructing the pignistic probability function in a context of\r\nuncertainty, Uncertainty in Artificial Intelligence, Vol. 5, pp. 29\u201339, Aug\r\n2004.\r\n[16] J. Sudano, Yet Another Paradigm Illustrating Evidence Fusion (YAPIEF),\r\nProc. of Fusion, Florence, July 2006.\r\n[17] T. L. Lai, Information bounds and quick detection of parameter changes\r\nin stochastic systems, in IEEE Transactions on Information Theory, vol.\r\n44, no. 7, pp. 2917\u20132929, Nov 1998.\r\n[18] I. F. Akyildiz, W. Lee, M. C. Vuran, and S. Mohanty, Next\r\ngeneration\/dynamic spectrum access\/cognitive radio wireless networks:\r\nA survey, Comput. Netw., vol. 50, no. 13, pp. 21272159, Sep. 2006.\r\n[19] S. Haykin, Cognitive Radio: Brain-Empowered Wireless\r\nCommunications, IEEE JSAC, vol.23, no.2, pp.201-20, Feb. 2005.\r\n[20] T. Yucek and H. Arslan, A survey of spectrum sensing algorithms\r\nfor cognitive radio applications, in IEEE Communications Surveys &\r\nTutorials, vol. 11, no. 1, pp. 116-130, First Quarter 2009.","publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 126, 2017"}