Abstract
The problem of determining the instants of an abrupt change in the statistical characteristics of a discrete-time binary Markovian process is considered. The problem solution method based on the Kullback-Leibler measure applied to estimation of the joint-probability distribution for different signal segments is proposed. It is shown that the use of the joint-probability distribution allows construction of a computationally efficient estimation algorithm.
Similar content being viewed by others
References
R. L. Stratonovich, Conditional Markovian Processes and Their Application to the Optimum Control Theory (Mosk. Gos. Univ., Moscow, 1966) [in Russian].
V. G. Repin, Probl. Peredachi Inf. 27(1), 61 (1991).
A. A. Mal’tsev and A. M. Silaev, Avtom. Telemekh., No. 1, 48 (1985).
S. Karlin, A First Course in Stochastic Processes (Academic, New York, 1966; Mir, Moscow, 1971) [in Russian].
V. I. Romanovskii, Discrete Markov Chains (Gostekhteorizdat, Moscow, 1949) [in Russian].
V. V. Yanshin, Radiotekh. Elektron. (Moscow) 38, 1081 (1993).
S. Kullback, Information Theory and Statistics (Wiley, New York, 1959; Nauka, Moscow, 1967).
C. Shannon, Works on Information Theory and Cybernetics (Inostrannaya Literatura, Moscow, 1963) [in Russian].
V. V. Zolotarev and G. V. Ovechkin, Error Correction Coding. Methods and Algorithms. Reference Book, Ed. by Yu. B. Zubarev (Goryachaya Liniya-Telekom, Moscow, 2004) [in Russian].
N. Merhav, M. Gutman, and J. Ziv, IEEE Trans. Inf. Theory 35, 1014 (1989).
Additional information
Original Russian Text © E.A. Kon’kov, O.A. Morozov, E.A. Soldatov, V.R. Fidel’man, 2007, published in Radiotekhnika i Elektronika, 2007, Vol. 52, No. 12 pp. 1458–1462.
Rights and permissions
About this article
Cite this article
Kon’kov, E.A., Morozov, O.A., Soldatov, E.A. et al. Application of the Kullback-Leibler measure for estimating the instants of a change in the statistical properties of a binary markovian process. J. Commun. Technol. Electron. 52, 1350–1354 (2007). https://doi.org/10.1134/S1064226907120078
Received:
Issue Date:
DOI: https://doi.org/10.1134/S1064226907120078