Application of Stein Rules to Combination Forecasting

We propose some Stein-rule combination forecasting methods that are designed to ameliorate the estimation risk inherent in making operational the variance–covariance method for constructing combination weights. By Monte Carlo simulation, it is shown that this amelioration can be substantial in many cases. Moreover, generalized Stein-rule combinations are proposed that offer the user the opportunity to enhance combination forecasting performance when shrinking the feasible variance–covariance weights toward a fortuitous shrinkage point. In an empirical exercise, the proposed Stein-rule combinations performed well relative to competing combination methods.     

Penalized least squares smoothing of two-dimensional mortality tables with imposed smoothness

This paper presents a method to estimate mortality trends of two-dimensional mortality tables. Comparability of mortality trends for two or more of such tables is enhanced by applying penalized least squares and imposing a desired percentage of smoothness to be attained by the trends. The smoothing procedure is basically determined by the smoothing parameters that are related to the percentage of smoothness. To quantify smoothness, we employ an index defined first for the one-dimensional case and then generalized to the two-dimensional one. The proposed method is applied to data from member countries of the OECD. We establish as goal the smoothed mortality surface for one of those countries and compare it with some other mortality surfaces smoothed with the same percentage of two-dimensional smoothness. Our aim is to be able to see whether convergence exists in the mortality trends of the countries under study, in both year and age dimensions.     

Efficiency of an adaptive estimator of a log-zero-poisson parameter

An estimator, λ is proposed for the parameter λ of the log-zero-Poisson distribution. While it is not a consistent estimator of λ in the usual statistical sense, it is shown to be quite close to the maximum likelihood estimates for many of the 35 sets of data on which it is tried. Since obtaining maximum likelihood estimates is extremely difficult for this and other contagious distributions, this estimate will act at least as an initial estimate in solving the likelihood equations iteratively. A lesson learned from this experience is that in the area of contagious distributions, variability is so large that attention should be focused directly on the mean squared error and not on consistency or unbiasedness, whether for small samples or for the asymptotic case. Sample sizes for some of the data considered in the paper are in hundreds. The fact that the estimator which is not a consistent estimator of λ is closer to the maximum likeli-hood estimator than the consistent moment estimator shows that the variability is large enough to not permit consistency to materialize even for such large sample sizes usually available in actual practice.     

A note on some conditions for optimality of certain ridge regression estimates

It is noted that the results obtained by Alam and Hawkes show that the ridge estimates suggested by Hoerl, Kennard and Baldwin, and by Lawless and Wang, give guaranteed lower mean square error than least squares for design matrices having none or at least two very small eigenvalues, but not necessarily for those having only one very small eigenvalue.     

Sequential point estimation of the means of u-statistics in finite population sampling

In simple random sampling without replacement from a finite population, sequential point estimators of the means of U-statistics are proposed. The proposed procedure is shown to be asymptotically risk efficient in the sense of Starr (Ann. Math. Statist. (1966), 1173-1185)     

Estimation of variance of a normal population

In this paper we have suggested two estimators of variance of a normal population developed from the estimators of u² suggested by Govindarajulu and Sahai and Das. These have been shown to be more efficient than the usual estimator s².     

Minimax linear empirical bayes estimation of the binomial parameter

The minimax linear Empirical Bayes estimators for a binomial parameter are obtained, assuming some information about the moments of the prior. The form of these estimates is used to propose a criterion which may be helpful in determining whether Empirical Bayes estimation is Indicated for a given problem.     

On shrinkage estimation of the exponential location parameter

Under the assumption that the exponential distribution is a reasonable model for a given population, some shrinkage estimators for the location parameter based on type 1 and type II censored samples have been derived. It is shown that these estimators dominate maximum likelihood estimators (MLE's) asymptotically under the mean squared error (MSE) criterion. A Monte Carlo study shows a significant improvement of our estimators over MLE's in terms of MSE for small samples.     

Estimation of error rate in multiple group logistic discrimination. the approximate leaving-one-out method

We present an approximate leaving-one-out technique for estimating the error rate in logistic discrimination. The new measure is based on the one-step approximation of a(i), the maximum likelihood estimate of the parameter vector based on the sample without the ith case. Some inequalities between the resubstitution error rate, the approximate and exact leaving-one-out error rates for the multiple group logistic model are investigated. Monte-Carlo simulations assess the adequacy of the approximate leaving-one-out method as an estimate of the actual error rate. The usefulness of this approach is demonstrated by means of two medical examples.     

Roundoff error for floating point representation of real data

In the present paper we define aprobabilistic model for the study ofthe relative roundoff error in the floating point representation ofreal data. In particular, starting from the assumption that real data is uniformly distributed in the interval (0,1), each floating point is assigned aprobability value corresponding to the proportion of real data assigned this floating point value. For each real number x～ U(0,1), we compute the relative roundoff error in the floating point representation and we calculate i t s expected value and variance.