Multiple hypotheses optimal testing for Markov chains and identification subject to the reliability criterion

• Autorzy: Navaei L.
• Języki publikacji: EN

Abstrakty

EN

The problem of many (L > 2)hypotheses testing on distributions of a finite state Markov chain is studied. We apply large deviation techniques (LDT). It is demonstrated that this method of investigation in solving the problem of logarithmically asymptotically optimal (LAO) hypotheses testing is easier, compared with the procedure introduced by Haroutunian. The matrix of exponents [formula/wzór], of error probabilities of the LAO test [formula/wzór] is the probability to accept the hypothesis l, when the hypothesis m is true, is determined. Moreover, the identification of distributions for one object and two independent objects via simple homogeneous stationary Markov chains with finite number of states is discussed in the present paper.

Słowa kluczowe

EN

error probability; irreducible matrix; stationary Markov chain; type

PL

prawdopodobieństwo błędu; macierz nierozkładalna; proces Markova; łańcuch Markova

• Wydawca: Oficyna Wydawnicza Politechniki Rzeszowskiej
• Czasopismo: Journal of Mathematics and Applications
• Rocznik: 2010
• Tom: Vol. 33
• Strony: 79--90
• Opis fizyczny: Bibliogr. 11 poz.

Twórcy

autor

Navaei L.

• Payame Noor University (P.N.U.), Faculty of Mathematics and Applied Probability, Iran,

Bibliografia

• [1] Ahlswede R. F. and Haroutunian E. A., On logarithmically asymptotically optimal testing of hypotheses and identi_cation. Lecture Notes in Computer Science, vol. 4123. General Theory of Information Transfer and Combinations, Springer, pp. 462-478, 2006.
• [2] Blahut R. E., Principles and Practice of Information Theory, reading, MA, Addison-wesley, 1987.
• [3] Csiszár I. and Shields P., Information Theory and Statistics, Fundations and Trends in Communications and Information Theory, vol. 1, no. 4, 2004.
• [4] Csiszár I. and Körner J., Information Theory: Coding Theorem for Discrete Memoryless Systems, Academic press, NewYork, 1981.
• [5] Dembo A. and Zeitouni O., Large Deviations Techniques and Applications, Jons and Bartlet. Publishers, London, 1993.
• [6] Gutman M., Asymptotically optimal classification for multiple test with empirically observed statistics, IEEE Trans. Inform. Theory, vol. 35, no. 2. pp. 401-408, 1989.
• [7] Haroutunian E. A., On asymptotically optimal testing of hypotheses concerning Markov chain, (in Russian). Izvestia Acad. Nauk Armenian SSR. Seria Mathem. vol. 22, no. 1. pp. 76-80, 1988.
• [8] Haroutunian E. A, Haroutunian M. E and Harutyunyan A. N., Reliability Criteria in Information Theory and in Statistical Hypothesis Testing, Foundations and Trends in Communications and Information Theory, vol. 4, no. 2-3, 2007.
• [9] Kullback S., Information Theory and Statistics, Wiley, New York, 1959.
• [10] Natarajan S., Large deviations, hypotheses testing, and source coding for finite Markov chain, IEEE Trans. Inform. Theory, vol. 31, no. 3, pp. 360-365, 1985.
• [11] Navaei L., Application of the theory of large deviations on error exponents in many hypotheses LAO testing, Journal of Statistics and Management Systems, New Delhi, vol. 11, no. 2, pp. 201-212, 2008.