Multivariate spatial U-quantiles: A Bahadur–Kiefer representation,a Theil–Sen estimator for multiple regression,and a robust dispersion estimator

A leading multivariate extension of the univariate quantiles is the so-called “spatial” or “geometric” notion, for which sample versions are highly robust and conveniently satisfy a Bahadur–Kiefer representation. Another extension of univariate quantiles has been to univariate U-quantiles, on the basis of which, for example, the well-known Hodges–Lehmann location estimator has a natural formulation. Generalizing both extensions, we introduce multivariate spatial U-quantiles and develop a corresponding Bahadur–Kiefer representation. New statistics based on spatial U-quantiles are presented for nonparametric estimation of multiple regression coefficients, extending the classical Theil–Sen nonparametric simple linear regression slope estimator, and for robust estimation of multivariate dispersion. Some other applications are mentioned as well.     

Orthogonal arrays and designs for partial triallel crosses

Abstract

Some block designs for Partial triallel crosses (PTC) are constructed in this paper using (a) orthogonal arrays (p², r, p, 2), where p is prime or a power of prime and (b) semi - balanced arrays (p (p???1)/2, p, p, 2), where p is a prime or power of an odd prime. Designs constructed are optimal in the sense of Kiefer (1975 Kiefer, J. 1975. Construction and optimality of generalized Youden designs. In A survey of statistical design and linear models, ed. J. N. Srivastava, 333–353. Amsterdam: North Hollond. [Google Scholar]). Some new designs for PTC experiments are also obtained.     

The Construction of Optimal Balanced Incomplete Block Designs When Adjacent Observations are Correlated

An algorithm is described for the optimal rearrangement of the treatments within each block of a Balanced Incomplete Block Design when a specified “nearest neighbour” correlation structure exists among observations from plots in the same block. The procedure uses results obtained by Kiefer & Wynn (1981). Designs obtained using the algorithm are found to compare favourably with those produced by combinatorial methods given in Cheng (1983). The algorithm produces optimal designs for all BIBD parameter sets, including those not covered by the results of Kiefer & Wynn or Cheng.     

On Non-parametric Testing, the Uniform Behaviour of the t-test, and Related Problems

Abstract.  In this article, we revisit some problems in non-parametric hypothesis testing. First, we extend the classical result of Bahadur & Savage [ Ann. Math. Statist . 25 (1956) 1115] to other testing problems, and we answer a conjecture of theirs. Other examples considered are testing whether or not the mean is rational, testing goodness-of-fit, and equivalence testing. Next, we discuss the uniform behaviour of the classical t -test. For most non-parametric models, the Bahadur–Savage result yields that the size of the t -test is one for every sample size. Even if we restrict attention to the family of symmetric distributions supported on a fixed compact set, the t -test is not even uniformly asymptotically level α . However, the convergence of the rejection probability is established uniformly over a large family with a very weak uniform integrability type of condition. Furthermore, under such a restriction, the t -test possesses an asymptotic maximin optimality property.     

A Direct Approach to Understanding Posterior Consistency of Bayesian Regression Problems

Previous approaches to establishing posterior consistency of Bayesian regression problems have used general theorems that involve verifying sufficient conditions for posterior consistency. In this article, we consider a direct approach by computing the posterior density explicitly and evaluating its asymptotic behavior. For this purpose, we deal with a sample size dependent prior based on a truncated regression function with increasing sample size, and evaluate the asymptotic properties of the resulting posterior. Based on a concept called posterior density consistency, we attempt to understand posterior consistency. As an application, we illustrate that the posterior density of an orthogonal semiparametric regression model is consistent.     

A note on the universal optimality criterion for full rank models

The aim of this note is to suggest a revised formulation of the universal optimality criterion for full rank models as stated in Kiefer (1975). We have presented the relevant results with indications of some possible applications.     

Long run variance estimation and robust regression testing using sharp origin kernels with no truncation

A new family of kernels is suggested for use in long run variance (LRV) estimation and robust regression testing. The kernels are constructed by taking powers of the Bartlett kernel and are intended to be used with no truncation (or bandwidth) parameter. As the power parameter (ρ)$(\rho )$ increases, the kernels become very sharp at the origin and increasingly downweight values away from the origin, thereby achieving effects similar to a bandwidth parameter. Sharp origin kernels can be used in regression testing in much the same way as conventional kernels with no truncation, as suggested in the work of Kiefer and Vogelsang [2002a, Heteroskedasticity-autocorrelation robust testing using bandwidth equal to sample size. Econometric Theory 18, 1350–1366, 2002b, Heteroskedasticity-autocorrelation robust standard errors using the Bartlett kernel without truncation, Econometrica 70, 2093–2095] Analysis and simulations indicate that sharp origin kernels lead to tests with improved size properties relative to conventional tests and better power properties than other tests using Bartlett and other conventional kernels without truncation.     

Optimal design for local fitting

This paper provides optimal design strategies for local (polynomial) fitting by the so-called moving local regression, a well-known nonparametric statistical tool. A Kiefer—Wolfowitz type equivalence theorem is formulated. Some examples are employed to illustrate the relations and differences to parametric techniques.     

On the asymptotics of randomness statistics

Kiefer (1959) studied the asymptotics of q-sample Cramér-von Mises nonparametric statistics when q is fixed and the sample sizes tend to infinity. Here we prove the asymptotic normality of such statistics when the sample sizes stay fixed or small while the number of samples, q, becomes large.     

COMPUTER AIDED-CONSTRUCTION OF D-OPTIMAL 2m FRACTIONAL DESIGNS OF RESOLUTION V

A new exchange algorithm for construction of 2^mD-optimal fractional factorial design (FFD) is devised. This exchange algorithm is a modification of the one due to Fedorov (1969, 1972) and is an improvement over similar algorithm due to Mitchell (1974) and Galil & Kiefer (1980). This exchange algorithm is then used to construct 54 D-optimal 2^m-FFD's of resolution V for m = 4,5,6.     

Product limit estimates: a generalized maximum likelihood study

The generalized maximum likelihood estimate (GMLE) assumptions are studied for four product-limit estimates (PLE): Censoring PLE (Kaplan-Meier estimate), truncation PLE, censoring-truncation PLE, and the degenerated PLE - the empirical distribution function. This paper shows that all the PLE's are also the GMLE's even if they are derived from partial likelihoods by natural parameterization techniques. However, a counter example is given to show that Kiefer Wolfowitz's assumption (1956) for consistency of GMLE can hardly be satisfied for un-dominated case.     

Some results on optimal design under a first-order autoregression and on finite williams type ii designs

It was shown, essentially, by Kiefer (1961) that the type II (a) design of Williams (1952) is asymptotically universally optimum for a first-order autoregression with parameter λ >0. We investigate any optimality properties these designs have when finite. We show that small differences in the definitions of the autoregression or of the design can lead to standard results in the theory of optiaml design no longer being applicable. We include some useful results on patterned matrices.     

Kiefer–Wolfowitz algorithm under quasi-associated random errors

In this article, we study the algorithm of Kiefer–Wolfowitz underquasi-associated random errors. We establish the complete convergence and obtain an exponential bound. Additionally, we build a confidence interval for the minimum. Numerical examples are sketched out to confirm the theoretical results and show the accuracy of the algorithm.     

LIMIT THEOREMS FOR ASYMPTOTICALLY MINIMAX ESTIMATION OF A DISTRIBUTION WITH INCREASING FAILURE RATE UNDER A RANDOM MIXED CENSORSHIP/TRUNCATION MODEL

ABSTRACT

The search for optimal non-parametric estimates of the cumulative distribution and hazard functions under order constraints inspired at least two earlier classic papers in mathematical statistics: those of Kiefer and Wolfowitz[1] Kiefer, J. and Wolfowitz, J. 1976. Asymptotically Minimax Estimation of Concave and Convex Distribution Functions. Z. Wahrsch. Verw. Gebiete, 34: 73–85. [Crossref], [Web of Science ®] , [Google Scholar] and Grenander[2] Grenander, U. 1956. On the Theory of Mortality Measurement. Part II. Scand. Aktuarietidskrift J., 39: 125–153.  [Google Scholar] respectively. In both cases, either the greatest convex minorant or the least concave majorant played a fundamental role. Based on Kiefer and Wolfowitz's work, Wang3-4 Wang, J.L. 1986. Asymptotically Minimax Estimators for Distributions with Increasing Failure Rate. Ann. Statist., 14: 1113–1131. Wang, J.L. 1987. Estimators of a Distribution Function with Increasing Failure Rate Average. J. Statist. Plann. Inference, 16: 415–427.   found asymptotically minimax estimates of the distribution function F and its cumulative hazard function Λ in the class of all increasing failure rate (IFR) and all increasing failure rate average (IFRA) distributions. In this paper, we will prove limit theorems which extend Wang's asymptotic results to the mixed censorship/truncation model as well as provide some other relevant results. The methods are illustrated on the Channing House data, originally analysed by Hyde.5-6 Hyde, J. 1977. Testing Survival Under Right Censoring and Left Truncation. Biometrika, 64: 225–230. Hyde, J. 1980. “Survival Analysis with Incomplete Observations”. In Biostatistics Casebook 31–46. New York: Wiley Series in Probability and Mathematical Statistics: Applied Probability and Statistics.       

Optimal designs for statistical analysis with Zernike polynomials

The Zernike polynomials arise in several applications such as optical metrology or image analysis on a circular domain. In the present paper, we determine optimal designs for regression models which are represented by expansions in terms of Zernike polynomials. We consider two estimation methods for the coefficients in these models and determine the corresponding optimal designs. The first one is the classical least squares method and Φ _p -optimal designs in the sense of Kiefer [Kiefer, J., 1974, General equivalence theory for optimum designs (approximate theory). Annals of Statistics, 2 849–879.] are derived, which minimize an appropriate functional of the covariance matrix of the least squares estimator. It is demonstrated that optimal designs with respect to Kiefer's Φ _p -criteria (p>?∞) are essentially unique and concentrate observations on certain circles in the experimental domain. E-optimal designs have the same structure but it is shown in several examples that these optimal designs are not necessarily uniquely determined. The second method is based on the direct estimation of the Fourier coefficients in the expansion of the expected response in terms of Zernike polynomials and optimal designs minimizing the trace of the covariance matrix of the corresponding estimator are determined. The designs are also compared with the uniform designs on a grid, which is commonly used in this context.     

On a cramér-von mises type statistic for testing bivariate independence

A Cramér-von Mises type statistic for testing bivariate independence, proposed by Hoeffding (1948) and by Blum, Kiefer, and Rosenblatt (1961), is examined in greater detail. The statistic is decomposed into components in the manner of Durbin and Knott (1972), and the components are shown to be related to linear rank statistics. Asymptotic power properties of the Hoeffding statistic and its components in testing for independence with bivariate normal random observations are described; a Monte Carlo study comparing these statistics with other nonparametric statistics for bivariate independence is also reported.     

Large and moderate deviations for modified Engel series

The modified Engel series of real numbers introduced by Rényi (1962 Rényi, A. (1962). A new approach to the theory of Engel’s series. Ann. Univ. Sci. Budapest. Eötvös Sect. Math. 5:25–32. [Google Scholar]) is a simple modification of Engel series, and they have the same classical limit theorems, such as the law of large numbers, central limit theorem, and law of the iterated logarithm. In this paper, we studied the large and moderate deviations for modified Engel series, which indicate that the large deviations for modified Engel series and Engel series are different.     

Inequality for equireplicated n-ary block designs with unequal block sizes

It is shown that certain inequalities known for binary, equireplicated, equiblock-sized block designs remain valid for equireplicated n-ary block designs with unequal block sizes. The approach used here is based on the spectral expansion of the C-matrix of the block design. The main theorems include some useful and combinatorially interesting results.     

Multidimensional strong large deviation theorems

In this paper we obtain some local limit theorems for arbitrary sequences of random vectors. The local limit theorems give conditions on the characteristic functions of random vectors for their pseudo-density function to converge uniformly on bounded sets. We then use these theorems to obtain strong large deviation results for an arbitrary sequence of random vectors. Thus our paper establishes the connection between the local limit theorems and the strong limit theorems. We apply our results to the multivariate F-distribution.