Comparison of families of algorithms for recognizing abnormal behavior of dynamic systems

详细信息    查看全文

• 作者：V. V. Shcherbinin (1)
  
  1. Department of Computational Mathematics and Cybernetics ; Moscow State University ; Moscow ; 119992 ; Russia
  
• 关键词：sample ; based learning problem ; parametric family of algorithms ; recognition algorithm ; multidimensional phase trajectory ; training sample
• 刊名：Moscow University Computational Mathematics and Cybernetics
• 出版年：2015
• 出版时间：January 2015
• 年：2015
• 卷：39
• 期：1
• 页码：41-47
• 全文大小：152 KB
• 参考文献：1. D. S. Kovalenko, Candidate鈥檚 Dissertation in Mathematical Physics (Mos. Gos. Univ., Moscow, 2011).
  2. Kostenko, V A, Shcherbinin, V V (2013) Training methods and algorithms for recognition of nonlinearly distorted phase trajectories of dynamic systems. Opt. Memory Neural Networks (Inf. Opt.) 22: pp. 8-20 CrossRef
  3. Ossovsky, S S (2002) Neural Networks for Information Processing. Finansy i statistika, Moscow
  4. Aivazyan, S A, Enyukov, I S, Meshalkin, L D (1989) Applied Statistics. Finansy i Statistika, Moscow
  5. Vorob鈥檈v, V I, Gribunin, V G (1999) Theory and Practice of Wavelet Transform. VUS, St. Petersburg
  6. Golyandina, N E, Nekrutkin, V V, Stepanov, D V (2003) Variants of the Caterpillar-SSA method for the analysis of multidimentional time series. System Identification and Control Problems (Proc. 2nd Int. Conf. SICPRO鈥?3, Moscow, 2003). IPU RAN, Moscow, pp. 2139-2168
  7. Muller, M (2007) Information Retrieval for Music and Motion. Springer-Verlag, New York CrossRef
  8. Kovalenko, D S, Kostenko, V A, Vasin, E A (2005) Study of the applicability of the algebraic approach to the analysis of time series. Methods and Means of Information Processing. VMiK Mos. Gos. Univ., Moscow, pp. 553-559
  9. Rudakov, K V, Chekhovich, Yu V (2001) On the problem of synthesis of learning trend selection algorithms (an algebraic approach). Applied Mathematics and Informatics. VMiK Mos. Gos. Univ., Moscow, pp. 97-114
  10. Kovalenko, D S, Kostenko, V A, Vasin, E A (2010) A genetic algorithm for construction of recognizers of anomalies in behaviour of dynamical systems. Proc. IEEE 5th Int. Conf. on Bio-inspired Computing: Theories and Applications, Changsha, China, 2010. IEEE, New York, pp. 258-263
  11. Schcherbinin, V V, Kostenko, V A (2013) A modification of training and recognition algorithms for recognition of abnormal behavior of dynamic systems. Proc. 5th Int. Joint Conf. on Computational Intelligence (IJCCI 2013), Vilamoura, Portugal, 2013. SciTePress, Portugal, pp. 103-110
  
• 刊物类别：Mathematics and Statistics
• 刊物主题：Mathematics
  Mathematics
  Russian Library of Science
  
• 出版者：Allerton Press, Inc. distributed exclusively by Springer Science+Business Media LLC
• ISSN：1934-8428

文摘

The problem of recognizing segments with abnormal behavior in multidimensional phase trajectories of dynamic systems is considered. Algorithms solving this problem with the help of the axiomatic approach to abnormal behavior recognition in dynamic systems are described. Two parametric families of recognition algorithms based on this approach are considered. For the first parametric family, a necessary condition for the existence of a recognition algorithm making a bounded number of errors on an arbitrary phase trajectory is proved. It is shown that this condition can be violated in the course of learning the first family, but this cannot take place with the second family. Results of a numerical study are presented that demonstrate that the second family provides a better recognition quality and requires less time for training an algorithm than the first family.