A class of modified stein estimators with easily computable risk functions

Consider the problem of estimating the mean of a p (≥3)-variate multi-normal distribution with identity variance-covariance matrix and with unweighted sum of squared error loss. A class of minimax, noncomparable (i.e. no estimate in the class dominates any other estimate in the class) estimates is proposed; the class contains rules dominating the simple James-Stein estimates. The estimates are essentially smoothed versions of the scaled, truncated James-Stein estimates studied by Efron and Morris. Explicit and analytically tractable expressions for their risks are obtained and are used to give guidelines for selecting estimates within the class.     

On a robustness property of PBIBD

In this paper we study a robustness property of partially balanced incomplete block designs based on association schemes with m classes (PBIBD(m)) against the unavailability of data in the sense that, when any t (a positive integer) observations are unavailable the design remains connected w.r.t. treatment. We characterize the robustness property of PBIBD(m) completely for m=2 and partially for m=3.     

The consistency of a matching test

Let $\text{X}{}^1{}_1\text{?X}{}^1{}_2\text{???X}{}^1{}_{\mathrm{n}}$ be the order statistics of a random sample from a distribution on [0, 1]. Let A_k, the kth match, be the event that $\text{X}{}^1{}_{\mathrm{k}}\text{?(}\text{(k?1)}\text{n}\text{k}\text{n}$], and let S_n be the total number of matches. The consistency of S_n for testing uniform df, U, against df G≠U is investigated, and it is shown that S_n is consistent if the intersection of G with U has Lebesgue measure zero. It is also consistent against a sequence of alternatives approaching U at a rate less faster than $\text{n}{}^{\mathrm{<mtext>-1</mtext><mtext>2</mtext>}}$.     

Stopping times of one-sample rank order sequential probability ratio tests

Weed, Bradley and Grovindarajulu (1974) propose one-sample probability ratio tests based on Lehmann alternatives. They also study the finite sure termination of the stopping times. Motivated by Stein's proof of (1946) of the termination of a sequential probability ratio test (SPRT) in the case of independent and identically distributed (i.i.d.) random variables and the work of Sethuraman (1970) for the two- sample rank order SPRT, we obtain a very mild condition (namely, that a certain random variable U(Z) is not identically zero) for the finite sure termination of the existence of the moment generating function (m.g.f.) for the stopping time of one- sample rank order SPRT's.     

Regular group divisible designs and Bhaskar Rao designs with block size three

Some recursive constructions are given for Bhaskar Rao designs. Using examples of these designs found by Shyam J. Singh, Rakesh Vyas and new ones given here we show the necessary conditions λ≡0 (mod 2), λυ(υ?1)≡0 (mod 24) are sufficient for the existence of Bhaskar Rao designs with one association class and block size 3. This result is used with a result of Street and Rodger to obtain regular partially balanced block designs with 2υ treatments, block size 3, λ₁=0, group size 2 and υ groups.     

A concise proof of a theorem on products of power sums

The heart of a direct solution of the problem of moments of moments is a formula for the expected value of a product of sample power sums in terms of a linear function of products of population power sums. Such a result was given by Skellam (1949) and, earlier and more generally, by Dwyer (1938). The approach here leads to a result for the easily handled case which features the product of unit power sums of each multivariate variable from which a general case comes by easy extension without the need of technical material required by the earlier approaches. A corollary provides a new solution for the problem of obtaining the unbiased estimate of any linear function of products of power sums.     

Factorial-balance and the analysis of designs with factorial structure

Cotter, John and Smith (1973) have given conditions for an incomplete block design to have orthogonal factorial structure. Further results on the intra-block analysis of such designs are given. The concept of balance in factorial design is discussed and results are given which enable the degree of balance in generalised cyclic designs to be determined.     

On-line control procedures under the random-walk model with measurement error and attribute observations

Assuming the quality characteristic process follows the random-walk model, the economic online control procedures are investigated by using the boundary correction technique when (1) the noise or measurement error follows either the random-walk model or the pure random model, and (2) the observations are taken by attributes only. Simple approximations for the long-run cost rate and optimal-control parameters are obtained.     

Efficiency of reduced logistic regression models

One feature of the usual polychotomous logistic regression model for categorical outcomes is that a covariate must be included in all the regression equations. If a covariate is not important in all of them, the procedure will estimate unnecessary parameters. More flexible approaches allow different subsets of covariates in different regressions. One alternative uses individualized regressions which express the polychotomous model as a series of dichotomous models. Another uses a model in which a reduced set of parameters is simultaneously estimated for all the regressions. Large-sample efficiencies of these procedures were compared in a variety of circumstances in which there was a common baseline category for the outcome and the covariates were normally distributed. For a correctly specified model, the reduced estimates were over 100% efficient for nonzero slope parameters and up to 500% efficient when the baseline frequency and the effect of interest were small. The individualized estimates could have efficiencies less than 50% when the effect of interest was large, but were also up to 130% efficient when the baseline frequency was large and the effect of interest was small. Efficiency was usually enhanced by correlation among the covariates. For an underspecified reduced model, asymptotic bias in the reduced estimates was approximately proportional to the magnitude of the omitted parameter and to the reciprocal of the baseline frequency.     

Maximum-likelihood estimation for the removal method

We prove that the profile log-likelihood function for the removal method of estimating population size is unimodal. The result is obtained by a variation-diminishing property of the Laplace transform. An implication of this result is that the likelihood-ratio confidence region for the population size is always an interval. Necessary and sufficient conditions for the existence of a finite maximum-likelihood estimator are presented. We also present evidence that the likelihood-ratio confidence interval for the population size has acceptable small-sample coverage properties.