A simple technique for proving unbiasedness of tests and confidence regions
A simple upper bound to the Bayes error probability for feature selection
In this paper, feature selection in multiclass cases for classification of remote-sensing images is addressed. A criterion based on a simple upper bound to the error probability of the Bayes classifier for the minimum error is proposed. This criterion has the advantage of selecting features having a link with the error probability with a low computational load. Experiments have been carried out in order to compare the performances provided by the proposed criterion with the ones of some of the widely...
A simplex design for gradient estimation in quadratic regression
A simplified treatment of the restricted analysis of a slightly disproportionate factorial experiment.
This paper considers a procedure to obtain effect estimators in the least squares analysis of a slightly disproportionate factorial design when a sample survey is made of the results of an extensive experiment. Explicit formulae have been found for the restricted estimators and their variances, when the constraints normally imposed upon a proportional model are used. In addition, an approximate analysis of the original model is used to perform that estimation, and an approximate analysis of variance...
A Special Representation of a Sufficient Randomization Kernel.
A spectral characterization of the behavior of discrete time AR–representations over a finite time interval
In this paper we investigate the behavior of the discrete time AR (Auto Regressive) representations over a finite time interval, in terms of the finite and infinite spectral structure of the polynomial matrix involved in the AR-equation. A boundary mapping equation and a closed formula for the determination of the solution, in terms of the boundary conditions, are also gived.
A square root filter for real time multivariate regression
A Stability Theorem for Statistical Experiments.
A statistical comparison of accelerated concrete testing methods.
A statistical comparison of three goodness-of-fit criteria used in modelling distances.
A statistical variance components framework for mapping imprinted quantitative trait locus in experimental crosses.
A stochastic Paris-Erdogan model for fatigue crack growth using two-state model.
A stronger version of matrix convexity as applied to functions of Hermitian matrices.
A study of the number of solutions of the system of the log-likelihood equations for the 3-parameter Weibull distribution
The maximum likelihood estimators of the parameters for the 3-parameter Weibull distribution do not always exist. Furthermore, computationally it is difficult to find all the solutions. Thus, the case of missing some solutions and among them the maximum likelihood estimators cannot be excluded. In this paper we provide a simple rule with help of which we are able to know if the system of the log-likelihood equations has even or odd number of solutions. It is a useful tool for the detection of all...
A study of the tangent space model of the von Mises-Fisher distrubution.
For a random rotation X = M0 eφ(ε) where M0 is a 3 x 3 rotation, ε is a trivariate random vector, and φ(ε) is a skew symmetric matrix, the least squares criterion consists of seeking a rotation M called the mean rotation minimizing tr[(M - E(X))t (M - E(X))]. Some conditions on the distribution of ε are set so that the least squares estimator is unbiased. Of interest is when ε is normally distributed N(0;Σ). Unbiasedness of the least squares estimator is dealt with according to eigenvalues of Σ.
A subcopula based dependence measure
A dependence measure for arbitrary type pairs of random variables is proposed and analyzed, which in the particular case where both random variables are continuous turns out to be a concordance measure. Also, a sample version of the proposed dependence measure based on the empirical subcopula is provided, along with an R package to perform the corresponding calculations.
A sufficient condition for admissibility in linear estimation
It was recently shown that all estimators which are locally best in the relative interior of the parameter set, together with their limits constitute a complete class in linear estimation, both unbiased and biased. However, not all these limits are admissible. A sufficient condition for admissibility of a limit was given by the author (1986) for the case of unbiased estimation in a linear model with the natural parameter space. This paper extends this result to the general linear model and to biased...
A sufficient statistic and a nonstandard linearization in nonlinear regression
A summary of block replacement policies
A summary of replacement models with changing failure distributions