摘要:

We obtain a tighter bound on the probability of making a correct selec tion in Paulson's procedure for selecting the normal population which has the largest population mean when the populations have a common unknown Tali ance. As a consequence, we are able to use a sharper value for the coristallt a*in Paulson's procedure. Simulation studies indicate that this leads to ari inlprovement in the expected number of stages to termination and expettetl total number of observations which is uniform in k, P*s and σ/δ*.

ISSN:0747-4946    

DOI:10.1080/07474949108836222

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

This study provides sufficient conditions under which the uniform distance between the distribution of random sums of weighted random variables and the normal distribution is of order n-a(log n)a+b, for α <1/2, n-1/2(log n)b+3/2, for b > -3/2, and n-1/2log2n for b = -3/2. Extensions to one-sided linear processes are also discussed.

ISSN:0747-4946    

DOI:10.1080/07474949108836223

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

In the presence of nuisance parameters, the Bhattacharyya type bound for the asymptotic variance of estimation procedures is obtained. It is shown that the modified maximum likelihood (ML) estimation procedures together with any stopping rule does not attain the bound. Further it is shown that the modified ML estimation procedure with the appropriate stopping rule is second order asymptotically efficient in some class of estimation procedures in the sense that it attains the lower bound for the asymptotic variance in the class.

ISSN:0747-4946    

DOI:10.1080/07474949108836224

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

A new class of stochastic approximation type estimators, different from those obtained via Robbins-Monro procedure, is introduced. Their asymptotic properties are studied. The proposed estimator is on the n-th step defined as the one step version M-estimator, where the estimator from the previous step is used as a preliminary one. This type of estimators is particularly useful in testing of constancy of the regression relationship over time.

ISSN:0747-4946    

DOI:10.1080/07474949108836225

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

In problems of optimal stopping for F((W(t)+x)+) -c(s + t) with W(t) denoting Brownian motion, bounds for the optimal stopping regions are derived which describe the growth of the optimal boundaries. This generalizes results by lrle (1988) where the case f(x) = x was treated. Extensions are given when positive part is replaced by absolute value and also when a random walk takes the place of Brownian motion.

ISSN:0747-4946    

DOI:10.1080/07474949108836226

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

This note concerns the estimation of treatment difference following a sequential test in which two treatments are to be compared and treatment allocation is data-dependent. Woodroofe(1989) considered this problem when the responses are independent and identically distributed. In this paper, it is shown that woodroof'es asymptotic results also work well for the modified sequential test of coad (1991), which allows for possible linear time trends in the data.

ISSN:0747-4946    

DOI:10.1080/07474949108836227

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

摘要:

Let W1,W2,...be independent and identically distributed, positive and continuous, random variables having all moments finite with E(W1) = 0 Suppose thatwhere. Let N1(h) be the smallest integerand define N(h) = max{t(h), Nl(h)}. We sample the difference {N(h) - t(h)} in one single batch and term N(h) as the accelerated stopping time. First we derive asymptotic (as h → 0) second-order properties of various positive and negative moments of N(h) under appropriate conditions on m. Then, this machinery is used for various statistical problems in constructing fixed-size confidence regions and minimum risk point estimators. For some of these problems, we discuss moderate sample size performances of N(h) and associated estimators obtained through simulated experiments and contrast with those of the corresponding customary purely sequential and three - stage estimation procedures. With "proper" choices of of ρ and q, we observe that the accelerated version N(h) can save considerable amount of sampling operations and its moderate sample properties are (i) remarkably similar to those of the purely sequential procedures, and (ii) often better than those of the three-stage procedures.

ISSN:0747-4946    

DOI:10.1080/07474949108836228

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor

ISSN:0747-4946    

DOI:10.1080/07474949108836221

出版商:Marcel Dekker, Inc.    

年代:1991

数据来源: Taylor