On the algebras of symmetries (groups of collineations) of designs from finite Desarguesian planes with applications in statistics

Decomposition into a direct sum of irreducible representations of the representation of the full collineation group of a finite Desarguesian plane, as a group of matrices permuting the flags of the plane and the simple components of the corresponding commutant algebra, have been worked out here for the projective plane PG(2, 2) and the affine plane EG(2, 3). The dimension and the components of the covariance matrix of the observations from a design derived from such a plane, which commutes with such a permutation representation of the full collineation group of the plane, are thus determined. This paper is in the spirit of earlier works by, James (1957), Mann (1960), 6., 7., McLaren (1963), and Sysoev and Shaikan (1976). A. T. James, Ann. Math. Statist.28 (1957), 993–1002, H. B. Mann, Ann. Math. Statist.31 (1960), 1–15, E. J. Hannan, Research Report (Part. (I)), Summer Research Institute, Australian Math. Soc. and Methuen's Monographs on Applied Probability and Statistics, Supplementary Review Series in Applied Probability, Vol. 3, A. D. McLaren, Proc. Cambridge Philos. Soc.59 (1963), 431–450, and L. P. Sysoev and M. E. Shaikin, Avtomat. i Telemekh.5 (1976), 64–73.     

On the robustness of least squares procedures in regression models

The criterion robustness of the standard likelihood ratio test (LRT) under the multivariate normal regression model and also the inference robustness of the same test under the univariate set up are established for certain nonnormal distributions of errors. Restricting attention to the normal distribution of errors in the context of univariate regression models, conditions on the design matrix are established under which the usual LRT of a linear hypothesis (under homoscedasticity of errors) remains valid if the errors have an intraclass covariance structure. The conditions hold in the case of some standard designs. The relevance of C. R. Rao's (1967 In Proceedings Fifth Berkeley Symposium on Math. Stat. and Prob., Vol. 1, pp. 355–372) and G. Zyskind's (1967, Ann. Math. Statist.38 1092–1110) conditions in this context is discussed.     

Admissible solutions of k-extended finite state set and sequence compound decision problems

We treat situations in which independent structurally identical decision problems are to be faced either simultaneously or serially. Recent work on such compound decision problems has centered on finding procedures that satisfy the strengthened asymptotic optimality property of Gilliland and Hannan (Ann. Math. Statist.40 (1969), 1536–1541) and on providing decision rules that at once satisfy an admissibility property and the classical asymptotic optimality property. We suggest a simplified method of accomplishing the first of these goals in general situations and then provide decision rules for finite state components simultaneously admissible and satisfying the strengthened optimality property.     

Elimination of randomization in statistical decision theory reconsidered

A result by Dvoretky, Wald, and Wolfowitz, (Ann. Math. Statist.22 (1951), 1–21) on the essential completeness of the class of nonrandomized deicision rules is generalized to a statistical decision model with noncompact decision space. The generalized result is obtained as a direct consequence of an existence result for an allocation problem arising in economics.     

LBI tests of independence in bivariate exponential distributions

The locally best invariant test for the hypothesis of independence in bivariate distributions with exponentially distributed marginals is derived. The model consists of a family of bivariate exponential distributions with probability density function $$f_\theta (x_1 ,x_2 ;\lambda _1 ,\lambda _2 ) = \lambda _1 \lambda _2 \exp [ - (\lambda _1 x_1 + \lambda _2 x_2 )]g(\lambda _1 x_1 ,\lambda _2 x_2 ;\theta )$$ with unknown scale parameter γ _j (j=1, 2) and association parameter ? which includes the independence situation. The locally best invariant (LBI) test is derived and the asymptotic null and nonnull distributions are also derived under some regularity conditions. The results are applied to the Gumbel (1960,J. Amer. Statist. Assoc.,55, 698–707), Frank (1979,Aequationes Math.,19, 194–226, and Cook and Johnson (1981),J. Roy. Statist. Soc. Ser. B,43, 210–218) families.     

Efficiency and Cramér-Rao type lnequalities for convex loss functions

A general method for obtaining inequalities of Cramér-Rao type for convex loss functions is presented. It is shown under rather weak assumptions that there are at least as many such inequalities as best unbiased estimators. More precisely, it is shown that an estimator is efficient with respect to an inequality of Cramér-Rao type if and only if it is the best in the class of unbiased estimators. Moreover, theorems of Blyth and Roberts (“Proceedings Sixth Berkeley Symposium on Math. Statist. Prob.,”) and of Blyth (Ann. Statist.2, 464–473) are extended. We make an use of methods of convex analysis and properties of convex integral functionals on Orlicz spaces.     

A Bahadur efficiency comparison between one and two sample rank statistics and their sequential rank statistic analogues

One and two sample rank statistics are shown in general to be more efficient in the Bahadur sense than their sequential rank statistic analogues as defined by Mason (1981, Ann. Statist.9 424–436) and Lombard (1981, South African Statist. J.15 129–152), even though the two families of statistics (those based on full ranks and those based on sequential ranks) have the same Pitman efficiency against local alternatives. In the process, general results on large deviation probabilities and laws of large numbers for statistics based on sequential ranks are obtained.     

A locally most powerful invariant test for the equality of means associated with covariate discriminant analysis

In this paper, the authors propose a locally most powerful invariant test for the equality of means in the presence of covariate variables. Also the null and nonnull distributions associated with the above test are developed. This problem arises in covariate discriminant analysis and has been treated by various authors, notably Cochran and Bliss 1948, Ann. Math. Statist.19, 151–176 and Rao 1949, Sankhyã9 343–366; 1966. The test derived here locally dominates in power the tests proposed so far. It is also shown that the Cochran-Bliss test is uniformly most powerful in the class of conditional invariant tests.     

Cross-validation and the smoothing of orthogonal series density estimators

We describe a class of smoothed orthogonal series density estimates, including the classical sequential-series introduced by [6], Soviet Math. Dokl. 3 1559–1562) and [16], Ann. Math. Statist. 38 1261–1265), and [23], Ann. Statist 9 146–156) two-parameter smoothing. The Bowman-Rudemo method of least-squares cross-validation (1982, Manchester-Sheffield School of Probability and Statistics Research Report 84/AWB/1; 1984, Biometrika 71 353–360; [14], Scand. J. Statist. 9 65–78), is suggested as a practical way of choosing smoothing parameters automatically. Using techniques of [18], Ann. Statist. 12 1285–1297), that method is shown to perform asymptotically optimally in the case of cosine and Hermite series estimators. The same argument may be used for other types of series.     

On Gronwall-Bellman-Bihari-type integral inequalities in several variables with retardation

Presented are some new nonlinear integral inequalities of the Gronwall-Bellman-Bihari type in n-independent variables with delay which extend recent results of C. C. Yeh and M.-H. Shin [J. Math. Anal. Appl.86, (1982), 157–167], C. C. Yeh [J. Math. Anal. Appl.87, (1982), 311–321], and A. I. Zahariev and D. D. Bainor [J. Math. Anal. Appl.89, (1981), 147–149]. Some applications of the results are included.     

Robust regression function estimation

A robust estimator of the regression function is proposed combining kernel methods as introduced for density estimation and robust location estimation techniques. Weak and strong consistency and asymptotic normality are shown under mild conditions on the kernel sequence. The asymptotic variance is a product from a factor depending only on the kernel and a factor similar to the asymptotic variance in robust estimation of location. The estimation is minimax robust in the sense of Huber (1964). Robust estimation of a location parameter. Ann. Math. Statist.33 73–101.     

Applications of autoreproducing kernel grammian moduli to S (U,H)-valued stationary random functions

It is shown that the analytical characterizations of q-variate interpolable and minimal stationary processes obtained by H. Salehi (Ark. Mat., 7 (1967), 305–311; Ark. Mat., 8 (1968), 1–6; J. Math. Anal. Appl., 25 (1969), 653–662), and later by A. Weron (Studia Math., 49 (1974), 165–183), can be easily extended to Hilbert space valued stationary processes when using the two grammian moduli that respectively autoreproduce their correlation kernel and their spectral measure. Furthermore, for these processes, a Wold-Cramér concordance theorem is obtained that generalizes an earlier result established by H. Salehi and J. K. Scheidt (J. Multivar. Anal., 2 (1972), 307–331) and by A. Makagon and A. Weron (J. Multivar. Anal., 6 (1976), 123–137).     

The infinitesimal Lefschetz formulas: A heat equation proof

In this paper, the Lefschetz formulas of Atiyah and Bott (Ann. of Math.86 (1967), 374–407; 88 (1968), 451–491) and Atiyah and Singer (Ann. of Math.87 (1968), 546–604; 93 (1971), 119–138) are proved by heat equation methods in infinitesimal form, i.e., without using the usual localization on the fixed point set. The proof is obtained by a homotopic deformation of the Dirac operator and uses probabilistic methods. The main applications are potentially in noncompact situations where fixed point formulas do not always hold, while their infinitesimal analogues are still valid.     

The entropy condition and the admissibility of shocks

The author proposed (Trans. Amer. Math. Soc.199 (1974), 89–112) the extended entropy condition (E) and solved the Riemann problem for general 2 × 2 conservation laws. The Riemann problem for 3 × 3 gas dynamics equations was treated by the author (J. Differential Equations18 (1975), 218–231). In this paper we justify condition (E) by the viscosity method in the spirit of Gelfand [Uspehi Mat. Nauk14 (1959), 87–158]. We show that a shock satisfies condition (E) if and only if the shock is admissible, that is, it is the limit of progressive wave solutions of the associated viscosity equations. For the “genuinely nonlinear” 2 × 2 conservation laws, Conley and Smoller [Comm. Pure Appl. Math.23 (1970), 867–884] proved that a shock satisfies Lax's shock inequalities [cf. Comm. Pure Appl. Math.14 (1957), 537–566] if and only if it is admissible. In this paper, we consider systems that are not necessarily genuinely nonlinear.     

Girth in graphs

It is shown that a graph of large girth and minimum degree at least 3 share many properties with a graph of large minimum degree. For example, it has a contraction containing a large complete graph, it contains a subgraph of large cyclic vertex-connectivity (a property which guarantees, e.g., that many prescribed independent edges are in a common cycle), it contains cycles of all even lengths modulo a prescribed natural number, and it contains many disjoint cycles of the same length. The analogous results for graphs of large minimum degree are due to Mader (Math. Ann.194 (1971), 295–312; Abh. Math. Sem. Univ. Hamburg37 (1972), 86–97), Woodall (J. Combin. Theory Ser. B22 (1977), 274–278), Bollobás (Bull. London Math. Soc.9 (1977), 97–98) and Häggkvist (Equicardinal disjoint cycles in sparse graphs, to appear). Also, a graph of large girth and minimum degree at least 3 has a cycle with many chords. An analogous result for graphs of chromatic number at least 4 has been announced by Voss (J. Combin. Theory Ser. B32 (1982), 264–285).     

An existence result for optimal economic growth problems

An existence result for optimal control problems of Lagrange type with unbounded time domain is derived very directly from a corresponding result for problems with bounded time domain. This subsumes the main existence result of R. F. Baum ¦J. Optim. Theory Appl.19 (1976), 89–116¦ and has the existence results for optimal economic growth problems of S.-I. Takekuma ¦J. Math. Econom.7 (1980), 193–208¦ and M. J. P. Magill ¦Econometrica49 (1981), 679–711; J. Math. Anal. Appl.82 (1981), 66–74¦ as simple corollaries. In addition, a new notion of uniform integrability is used, which coincides with the classical notion if the time domain is bounded.     

A new extreme quantile estimator for heavy-tailed distributions

The classical estimation method for extreme quantiles of heavy-tailed distributions was presented by Weissman (J. Amer. Statist. Assoc. 73 (1978) 812–815) and makes use of the Hill estimator (Ann. Statist. 3 (1975) 1163–1174) for the positive extreme value index. This index estimator can be interpreted as an estimator of the slope in the Pareto quantile plot in case one considers regression lines passing through a fixed anchor point. In this Note we propose a new extreme quantile estimator based on an unconstrained least squares estimator of the index, introduced by Kratz and Resnick (Comm. Statist. Stochastic Models 12 (1996) 699–724) and Schultze and Steinebach (Statist. Decisions 14 (1996) 353–372) and we study its asymptotic behavior. To cite this article: A. Fils, A. Guillou, C. R. Acad. Sci. Paris, Ser. I 338 (2004).     

On the commutant algebras corresponding to the permutation representations of the full collineation groups of PG(k,s) and EG(k,s), s = pr,k ⩾ 2

In this paper, the dimension t and a linear basis of the commutant algebra corresponding to the representation of the full collineation group as matrices permuting the flags (incident point-line or point-hyperplane pairs) have been determined for each one of the four geometries PG(2, s), EG(2, s), PG(k, s), and EG(k, s), s = p^r, k ? 3. For the four geometries, t = 6, 7, 7, and 8, respectively, and the corresponding linear bases are (I, G, B, T, BT, TB), (I, G, B, T, BT, TB, BTB), (I, G, B, T, BT, TB, S), and (I, G, B, T, BT, TB, BTB, S). I, G, B, T are the relationship matrices of James (Ann. Math. Statist.28 (1957), 993–1082) and the matrix S was introduced by Sysoev and Shaikin (Avtomat. i Telemekh.5 (1976), 64–73).     

ε-Forced adaptive policy in age replacement

We consider the problem of minimizing the long-run average cost per unit time in age replacement with an unknown failure distribution. Utilizing the idea of ε-forced choices, which may be thought of as a modification of forced choice circles given by Fox and Rolph (Adaptive policies for Markov renewal programs, Ann. Statist.50 (1973), 334–341), we construct the adaptive policy which has a non-parametrically good property and is Bayes against a Dirichlet process prior.     

The capacity and equivocation of a transducer and a connection with Billingsley dimension

In this paper the study of a noiseless transducer with finite memory, begun by Conner [Ann. Math. Statist.41, No. 6 (1970)], is set forth. The uniqueness and the weak Bernoulli property of the maximal-output dynamical system are proved. A convergence theorem about equivocation, which is analogous to McMillan's theorem about entropy, is obtained. An application of this theorem yields a lower estimation of the Billingsley dimension of the ambiguity set of almost all output messages.