Large Sample Change-Point Estimation when Distributions Are Unknown

A. A. Borovkov

Teoriâ veroâtnostej i ee primeneniâ, Tome 53 (2008) no. 3, pp. 437-457

Résumé

Let $\textrm{X}=(\textrm{x}_1,\textrm{x}_2,\dots,\textrm{x}_n)$ be a sample consisting of $n$ independent observations in an arbitrary measurable space $\mathscr{X}$ such that the first $\theta$ observations have a distribution $F$ while the remaining $n-\theta$ ones follow $G\neq F$, the distributions $F$ and $G$ being unknown and quantities $n$ and $\theta$ large. In [A. A. Borovkov and Yu. Yu. Linke, Math. Methods Statist., 14 (2005), pp. 404–430] there were constructed estimators $\theta^*$ of the change-point $\theta$ that have proper error (i.e., such that $P_\theta\{|\theta^*-\theta|>k\}$ tends to zero as $k$ grows to infinity), under the assumption that we know a function $h$ for which the mean values of $h(\textrm{x}_j)$ under the distributions $F$ and $G$ are different from each other. Sequential procedures were also presented in that paper. In the present paper, we obtain similar results under a weakened form of the above assumption or even in its absence. One such weaker version assumes that we have functions $h_1,h_2,\ldots,h_l$ on $\mathscr{X}$ such that for at least one of them the mean values of $h_j(\textrm{x}_i)$ are different under $F$ and $G$. Another version does not assume the existence of known to us functions $h_j$, but allows the possibility of estimating the unknown distributions $F$ and $G$ from the initial and terminal segments of the sample $\textrm{X}$. Sequential procedures are also dealt with.

Détail

Keywords: change-point problem for unknown distributions, change-point, sequential estimation.

Bibliographie

[1] Borovkov A. A., Linke Yu. Yu., “Change-point problem for large samples and incomplete information on distributions”, Math. Methods Statist., 14:4 (2005), 404–430 | MR

[2] Darkhovskii B. S., “Neparametricheskii metod dlya aposteriornogo obnaruzheniya momenta “razladki” posledovatelnosti nezavisimykh sluchaiykh velichin”, Teoriya veroyatn. i ee primen., 21:1 (1976), 180–184 | MR

[3] Carlstein E., “Nonparametric change-point estimation”, Ann. Statist., 16:1 (1988), 188–197 | DOI | MR | Zbl

[4] Hawkins D. L., “A simple least squares method for estimating a change in mean”, Comm. Statist. Simulation, 15 (1986), 655–679 | MR | Zbl

[5] Dumbgen L., “The asymptotic behavior of some nonparametric change-point estimators”, Ann. Statist., 19:3 (1991), 1471–1495 | DOI | MR

[6] Boukai B., Zhou H., “Nonparametric estimation in a two change-point model”, J. Nonparametr. Statist., 8:3 (1997), 275–292 | DOI | MR | Zbl

[7] Huang W.-T., Chang Y.-P., “Nonparametric estimation in change-point models”, J. Statist. Plann. Inference, 35:3 (1993), 335–347 | DOI | MR | Zbl

[8] Baron M. I., “Nonparametric adaptive change-point estimation and on-line detection”, Sequential Anal., 19:1–2 (2000), 1–23 | DOI | MR | Zbl

[9] Brodsky B. E., Darkhovsky B. S., Nonparametric Methods in Change-point Problems, Kluwer, Dordrecht, 1993, 209 pp. | MR | Zbl

[10] Borovkov A. A., “Asimptoticheski optimalnye resheniya v zadache o razladke”, Teoriya veroyatn. i ee primen., 43:4 (1998), 625–654 | MR | Zbl

[11] Borovkov A. A., Matematicheskaya statistika, Nauka, Novosibirsk, 1997, 771 pp. | MR

[12] Borovkov A. A., Veroyatnostnye protsessy v teorii massovogo obsluzhivaniya, Nauka, M., 1972, 367 pp. | MR

[13] Massart P., “The tight constant in the Dvoretzky–Kiefer–Wolfowitz inequality”, Ann. Probab., 18:3 (1990), 1269–1283 | DOI | MR | Zbl

[14] Borovkov A. A., “K zadache o dvukh vyborkakh”, Izv. AN SSSR, 26:4 (1962), 605–624 | MR | Zbl

[15] Sanov I. N., “O veroyatnosti bolshikh uklonenii sluchainykh velichin”, Matem. sb., 42:1 (1957), 11–44 | MR | Zbl

[16] Borovkov A. A., “Granichnye zadachi dlya sluchainykh bluzhdanii i bolshie ukloneniya v funktsionalnykh prostranstvakh”, Teoriya veroyatn. i ee primen., 12:4 (1967), 635–654 | MR | Zbl