On optimality of the Shiryaev-Roberts procedure for detecting a change in distribution. (English) Zbl 1204.62141

Polunchenko, Aleksey S.; Tartakovsky, Alexander G.

Ann. Stat. 38, No. 6, 3445-3457 (2010).

Summary: For detecting a change in distribution, M. Pollak [Ann. Stat. 13, 206–227 (1985; Zbl 0573.62074)] introduced a specific minimax performance metric and a randomized version of the Shiryaev-Roberts procedure [A. N. Shiryaev, Theor. Probab. Appl. 8, 22–46 (1963); translation from Teor. Veroyatn. Primen. 8, 26–51 (1963; Zbl 0213.43804); S. W. Roberts, Technometrics 8, 411–430 (1966)] where the zero initial condition is replaced by a random variable sampled from the quasi-stationary distribution of the Shiryaev-Roberts statistic. Pollak proved that this procedure is third-order asymptotically optimal as the mean time to false alarm becomes large. The question of whether Pollak’s procedure is strictly minimax for any false alarm rate has been open for more than two decades, and there were several attempts to prove this strict optimality. We provide a counterexample which shows that Pollak’s procedure is not optimal and that there is a strictly optimal procedure which is nothing but the Shiryaev-Roberts procedure that starts with a specially designed deterministic point.

MSC:

62L10	Sequential statistical analysis
62L15	Optimal stopping in statistics
60G40	Stopping times; optimal stopping problems; gambling theory

Keywords:

changepoint problems; Shiryaev-Roberts procedures; sequential detection

Citations:

Zbl 0573.62074; Zbl 0213.43804

References:

[1]	Feinberg, E. A. and Shiryaev, A. N. (2006). Quickest detection of drift change for Brownian motion in generalized Bayesian and minimax settings. Statist. Decisions 24 445-470. · Zbl 1135.60024
[2]	Harris, T. E. (1963). The Theory of Branching Processes . Springer, Berlin. · Zbl 0117.13002
[3]	Mei, Y. (2006). Comments on, “A note on optimal detection of a change in distribution,” by Benjamin Yakir. Ann. Statist. 34 1570-1576. · Zbl 1113.62092
[4]	Moustakides, G. V., Polunchenko, A. S. and Tartakovsky, A. G. (2009). Numerical comparison of CUSUM and Shiryaev-Roberts procedures for detecting changes in distributions. Comm. Statist. Theory Methods 38 3225-3239. · Zbl 1175.62084
[5]	Moustakides, G. V., Polunchenko, A. S. and Tartakovsky, A. G. (2011). A numerical approach to comparative efficiency analysis of quickest change-point detection procedures. Statist. Sinica . 21 . · Zbl 1214.62084
[6]	Pollak, M. (1985). Optimal detection of a change in distribution. Ann. Statist. 13 206-227. · Zbl 0573.62074
[7]	Pollak, M. (1987). Average run lengths of an optimal method of detecting a change in distribution. Ann. Statist. 15 749-779. · Zbl 0632.62080
[8]	Pollak, M. and Tartakovsky, A. G. (2009). Optimality properties of the Shiryaev-Roberts procedure. Statist. Sinica 19 1729-1739. · Zbl 05629283
[9]	Roberts, S. W. (1966). A comparison of some control chart procedures. Technometrics 8 411-430. JSTOR:
[10]	Shiryaev, A. N. (1961). The problem of the most rapid detection of a disturbance in a stationary process. Dokl. Math. 2 795-799. · Zbl 0109.11201
[11]	Shiryaev, A. N. (1963). On optimum methods in quickest detection problems. Theory Probab. Appl. 8 22-46. · Zbl 0213.43804
[12]	Yakir, B. (1997). A note on optimal detection of a change in distribution. Ann. Statist. 25 2117-2126. · Zbl 0942.62088