On the statistical computation of the sample multiple correlation moefficient

Based on the recursive formulas of Lee (1988) and Singh and Relyea (1992) for computing the noncentral F distribution, a numerical algorithm for evaluating the distributional values of the sample squared multiple correlation coefficient is proposed. The distributional function of this statistic is usually represented as an infinite weighted sum of the iterative form of incomplete beta integral. So an effective algorithm for the incomplete beta integral is crucial to the numerical evaluation of various distribution values. Let a and b denote two shape parameters shown in the incomplete beta integral and hence formed in the sampling distribution functionn be the sample size, and p be the number of random variates. Then both 2a = p - 1 and 2b = n - p are positive integers in sampling situations so that the proposed numerical procedures in this paper are greatly simplified by recursively formulating the incomplete beta integral. By doing this, it can jointly compute the distributional values of probability dens function (pdf) and cumulative distribution function (cdf) for which the distributional value of quantile can be more efficiently obtained by Newton's method. In addition, computer codes in C are developed for demonstration and performance evaluation. For the less precision required, the implemented method can achieve the exact value with respect to the jnite significant digit desired. In general, the numerical results are apparently better than those by various approximations and interpolations of Gurland and Asiribo (1991),Gurland and Milton (1970), and Lee (1971, 1972). When b = (1/2)(n -p) is an integer in particular, the finite series formulation of Gurland (1968) is used to evaluate the pdf/cdf values without truncation errors, which are served as the pivotal one. By setting the implemented codes with double precisions, the infinite series form of derived method can achieve the pivotal values for almost all cases under study. Related comparisons and illustrations are also presented     

Numerical algorithms for solving nonlinear L р-norm estimation problems: part II - a mixture method for large residual and illo-conditioned problems

The nonlinear least squares algorithm of Gill and Murray (1978) is extended and modified to solve nonlinear L _р-norm estimation problems efficiently. The new algorithm uses a mixture of 1st-order derivative (Guass-Newton) and 2nd-order derivative (Newton) search directions. A new rule for selecting the “grade” r of the p-jacobiab matrix J_p was also incorporated. This brought about rapid convergence of the algorithm on previously reported test examples.     

Balanced treatment incomplete block designs through integer programming

An algorithm is presented to construct balanced treatment incomplete block (BTIB) designs using a linear integer programming approach. Construction of BTIB designs using the proposed approach is illustrated with an example. A list of efficient BTIB designs for 2 ? v ? 12, v + 1 ? b ? 50, 2 ? k ? min(10, v), r ? 10, r₀ ? 20 is provided. The proposed algorithm is implemented as part of an R package.     

Aggregating classifiers via Rademacher–Walsh polynomials

Methods are proposed to combine several individual classifiers in order to develop more accurate classification rules. The proposed algorithm uses Rademacher–Walsh polynomials to combine M (≥2) individual classifiers in a nonlinear way. The resulting classifier is optimal in the sense that its misclassification error rate is always less than, or equal to, that of each constituent classifier. A number of numerical examples (based on both real and simulated data) are also given. These examples demonstrate some new, and far-reaching, benefits of working with combined classifiers.     

The optimal choice of negative binomial charts for monitoring high-quality processes

Good control charts for high quality processes are often based on the number of successes between failures. Geometric charts are simplest in this respect, but slow in recognizing moderately increased failure rates p. Improvement can be achieved by waiting until r>1 failures have occurred, i.e. by using negative binomial charts. In this paper we analyze such charts in some detail. On the basis of a fair comparison, we demonstrate how the optimal r is related to the degree of increase of p. As in practice p will usually be unknown, we also analyze the estimated version of the charts. In particular, simple corrections are derived to control the nonnegligible effects of this estimation step.     

A Property of Count Distributions in the Hinde–Demétrio Family

The Hinde–Demétrio (HD) family of distributions, which are discrete exponential dispersion models with an additional real index parameter p, have been recently characterized from the unit variance function μ + μ ^p . For p equals to 2, 3,…, the corresponding distributions are concentrated on non negative integers, overdispersed and zero-inflated with respect to a Poisson distribution having the same mean. The negative binomial (p = 2) and strict arcsine (p = 3) distributions are HD families; the limit case (p → ∞) is associated to a suitable Poisson distribution. Apart from these count distributions, none of the HD distributions has explicit probability mass functions p _k . This article shows that the ratios r _k  = k p _k /p _k?1, k = 1,…, p ? 1, are equal and different from r _p . This new property allows, for a given count data set, to determine the integer p by some tests. The extreme situation of p = 2 is of general interest for count data. Some examples are used for illustrations and discussions.     

Stochastic matching pursuit for Bayesian variable selection

This article proposes a stochastic version of the matching pursuit algorithm for Bayesian variable selection in linear regression. In the Bayesian formulation, the prior distribution of each regression coefficient is assumed to be a mixture of a point mass at 0 and a normal distribution with zero mean and a large variance. The proposed stochastic matching pursuit algorithm is designed for sampling from the posterior distribution of the coefficients for the purpose of variable selection. The proposed algorithm can be considered a modification of the componentwise Gibbs sampler. In the componentwise Gibbs sampler, the variables are visited by a random or a systematic scan. In the stochastic matching pursuit algorithm, the variables that better align with the current residual vector are given higher probabilities of being visited. The proposed algorithm combines the efficiency of the matching pursuit algorithm and the Bayesian formulation with well defined prior distributions on coefficients. Several simulated examples of small n and large p are used to illustrate the algorithm. These examples show that the algorithm is efficient for screening and selecting variables.     

Optimal component test plans for a parallel system based on Type-II censoring

Consider a parallel system with n independent components. Assume that the lifetime of the jth component follows an exponential distribution with a constant but unknown parameter λ_j, 1≤j≤n. We test r_j components of type-j for failure and compute the total time T_j of r_j failures for the jth component. Based on T=(T₁,T₂,…,T_n) and r=(r₁,r₂,…,r_n), we derive optimal reliability test plans which ensure the usual probability requirements on system reliability. Further, we solve the associated nonlinear integer programming problem by a simple enumeration of integers over the feasible range. An algorithm is developed to obtain integer solutions with minimum cost. Finally, some examples have been discussed for various levels of producer’s and consumer’s risk to illustrate the approach. Our optimal plans lead to considerable savings in costs over the available plans in the literature.     

Regularized proportional odds models

The proportional odds model (POM) is commonly used in regression analysis to predict the outcome for an ordinal response variable. The maximum likelihood estimation (MLE) approach is typically used to obtain the parameter estimates. The likelihood estimates do not exist when the number of parameters, p, is greater than the number of observations n. The MLE also does not exist if there are no overlapping observations in the data. In a situation where the number of parameters is less than the sample size but p is approaching to n, the likelihood estimates may not exist, and if they exist they may have quite large standard errors. An estimation method is proposed to address the last two issues, i.e. complete separation and the case when p approaches n, but not the case when p>n. The proposed method does not use any penalty term but uses pseudo-observations to regularize the observed responses by downgrading their effect so that they become close to the underlying probabilities. The estimates can be computed easily with all commonly used statistical packages supporting the fitting of POMs with weights. Estimates are compared with MLE in a simulation study and an application to the real data.     

Deficiencies of minimum discrepancy estimators

Suppose the multinomial parameters p_r (θ) are functions of a real valued parameter 0, r = 1,2, …, k. A minimum discrepancy (m.d.) estimator θ of θ is defined as one which minimises the discrepancy function D = Σ n_rf(p_r/n_r), for a suitable function f where n_r is the relative frequency in r-th cell, r = 1,2, …, k. All the usual estimators like maximum likelihood (m. l), minimum chi-square (m. c. s.)., etc. are m.d. estimators. All m.d. estimators have the same asymptotic (first order) efficiency. They are compared on the basis of their deficiencies, a concept recently introduced by Hodges and Lehmann [2]. The expression for least deficiency at any θ is derived. It is shown that in general uniformly least deficient estimators do not exist. Necessary and sufficient conditions on p_r (0) for m. t. and m. c. s. estimators to be uniformly least deficient are obtained.     

An optimal sign test for one-sample bivariate location model using an alternative bivariate ranked-set sample

The aim of this paper is to find an optimal alternative bivariate ranked-set sample for one-sample location model bivariate sign test. Our numerical and theoretical results indicated that the optimal designs for the bivariate sign test are the alternative designs with quantifying order statistics with labels {((r+1)/2, (r+1)/2)}, when the set size r is odd and {(r/2+1, r/2), (r/2, r/2+1)} when the set size r is even. The asymptotic distribution and Pitman efficiencies of these designs are derived. A simulation study is conducted to investigate the power of the proposed optimal designs. Illustration using real data with the Bootstrap algorithm for P-value estimation is used.     

Performance characteristics of the adjusted r2 algorithm for determining the start of the terminal disposition phase and comparison with a simple r2 algorithm and a visual inspection method

The adjusted r² algorithm is a popular automated method for selecting the start time of the terminal disposition phase (t_z) when conducting a noncompartmental pharmacokinetic data analysis. Using simulated data, the performance of the algorithm was assessed in relation to the ratio of the slopes of the preterminal and terminal disposition phases, the point of intercept of the terminal disposition phase with the preterminal disposition phase, the length of the terminal disposition phase captured in the concentration‐time profile, the number of data points present in the terminal disposition phase, and the level of variability in concentration measurement. The adjusted r² algorithm was unable to identify t_z accurately when there were more than three data points present in a profile's terminal disposition phase. The terminal disposition phase rate constant (λ_z) calculated based on the value of t_z selected by the algorithm had a positive bias in all simulation data conditions. Tolerable levels of bias (median bias less than 5%) were achieved under conditions of low measurement variability. When measurement variability was high, tolerable levels of bias were attained only when the terminal phase time span was 4 multiples of t_1/2 or longer. A comparison of the performance of the adjusted r² algorithm, a simple r² algorithm, and t_z selection by visual inspection was conducted using a subset of the simulation data. In the comparison, the simple r² algorithm performed as well as the adjusted r² algorithm and the visual inspection method outperformed both algorithms. Recommendations concerning the use of the various t_z selection methods are presented.     

Exact p Values for F and t Tests

A method is established for finding exact p values for F and t tests with the aid of a hand calculator. The derivation of the method is an exercise in probability theory and arithmetic series. With this method, the number of steps to a solution is always known. Numerical examples are given and comments and suggestions are made.     

An optimization approach applied to the optimal stratification problem

A new optimization algorithm is presented to solve the stratification problem. Assuming the number L of strata and the total sample size n are fixed, we obtain strata boundaries by using an objective function associated with the variance. In this problem, strata boundaries must be determined so that the elements in each stratum are more homogeneous among themselves. To produce more homogeneous strata, this paper proposes a new algorithm that uses the Greedy Randomized Adaptive Search Procedure (GRASP) methodology. Computational results are presented for a set of problems, with the application of the new algorithm and some algorithms from literature.     

Some small-sample properties of some recently proposed multivariate outlier detection techniques

Recently, several new robust multivariate estimators of location and scatter have been proposed that provide new and improved methods for detecting multivariate outliers. But for small sample sizes, there are no results on how these new multivariate outlier detection techniques compare in terms of p _n , their outside rate per observation (the expected proportion of points declared outliers) under normality. And there are no results comparing their ability to detect truly unusual points based on the model that generated the data. Moreover, there are no results comparing these methods to two fairly new techniques that do not rely on some robust covariance matrix. It is found that for an approach based on the orthogonal Gnanadesikan–Kettenring estimator, p _n can be very unsatisfactory with small sample sizes, but a simple modification gives much more satisfactory results. Similar problems were found when using the median ball algorithm, but a modification proved to be unsatisfactory. The translated-biweights (TBS) estimator generally performs well with a sample size of n≥20 and when dealing with p-variate data where p≤5. But with p=8 it can be unsatisfactory, even with n=200. A projection method as well the minimum generalized variance method generally perform best, but with p≤5 conditions where the TBS method is preferable are described. In terms of detecting truly unusual points, the methods can differ substantially depending on where the outliers happen to be, the number of outliers present, and the correlations among the variables.     

Bayesian one-sample prediction of future observations under Pareto distribution

Suppose that the length of time in years for which a business operates until failure has a Pareto distribution. Let t ₁?t ₂?t _r denote the survival lifetimes of the first r of a random sample of n businesses. Bayesian predictions are to be made on the ordered failure times of the remaining (n???r) businesses, using the conditional probability function. Numerical examples are given to illustrate our results.     

Fitting the additive model by recursion on dimension

We consider estimation for the homoscedastic additive model for multiple regression. A recursion is proposed in Opsomer (1999), and independently by the authors, for obtaining the estimators that solve the normal equations given by Hastie and Tibshirani (1990). The recursion can be exploited to obtain the asymptotic bias and variance expressions of the estimators for any p > 2 (Opsomer 1999) using repeated application of Opsomer and Ruppert (1997). Opsomer and Ruppert (1997) provide asymptotic bias and variance for the estimators when p = 2. Opsomer (1999) also uses the recursion to provide sufficient conditions for convergence of the backfitting algorithm to a unique solution of the normal equations. However, since explicit expressions for the solution to the normal equations are not given, he states, “The lemma does not provide a practical way of evaluating the existence and uniqueness of the backfitting estimators … ”. In this paper, explicit expressions for the estimators are derived. The explicit solution requires inverses of n × n matrices to solve the np × np system of normal equations. These matrix inverses are feasible to implement for moderate sample sizes and can be used in place of the backfitting algorithm.     

Csiszar’s ϕ-divergences for testing the order in a Markov chain

x ₁, ..., x _n+r can be treated as the sample values of a Markov chain of order r or less (chain in which the dependence extends over r+1 consecutive variables only), and consider the problem of testing the hypothesis H ₀ that a chain of order r− 1 will be sufficient on the basis of the tools given by the Statistical Information Theory: _ϕ-Divergences. More precisely, if p _{a 1 ....., a r}: _{a r +1} denotes the transition probability for a r ^th order Markov chain, the hypothesis to be tested is H ₀:p _{a 1 ....., a r}: _{a r +1} = p _{a 2 ....., a r}: _{a r +1}, a _i ∈{1, ..., s}, i = 1, ..., r + 1 The tests given in this paper, for the first time, will have as a particular case the likelihood ratio test and the test based on the chi-squared statistic. Received: August 3, 1998; revised version: November 25, 1999     

On the maxima of heterogeneous gamma variables

In this article, we establish some new results on stochastic comparisons of the maxima of two heterogenous gamma variables with different shape and scale parameters. Let X₁ and X₂ [X*₁ and X*₂] be two independent gamma variables with X_i?[X*_i] having shape parameter r_i?[r*_i] and scale parameter λ_i?[λ*_i], i = 1, 2. It is shown that the likelihood ratio order holds between the maxima, X_{2: 2} and X*_{2: 2} when λ₁ = λ*₁ ? λ₂ = λ*₂ and r₁ ? r*₁ ? r₂ = r*₂. We also prove that, if r_i, r*_i ∈ (0, 1], (r₁, r₂) majorizes (r*₁, r₂*), and (λ₁, λ₂) is p-larger than (λ*₁, λ₂*), then X_{2: 2} is larger than X*_{2: 2} in the sense of the hazard rate order [dispersive order]. Some numerical examples are provided to illustrate the main results. The new results established here strengthen and generalize some of the results known in the literature.     

On bootstrapping the number of components in finite mixtures of Poisson distributions

Finite mixture models arise in a natural way in that they are modeling unobserved population heterogeneity. It is assumed that the population consists of an unknown number k of subpopulations with parameters λ₁, ..., λ_k receiving weights p₁, ..., p_k. Because of the irregularity of the parameter space, the log-likelihood-ratio statistic (LRS) does not have a (χ²) limit distribution and therefore it is difficult to use the LRS to test for the number of components. These problems are circumvented by using the nonparametric bootstrap such that the mixture algorithm is applied B times to bootstrap samples obtained from the original sample with replacement. The number of components k is obtained as the mode of the bootstrap distribution of k. This approach is presented using the Times newspaper data and investigated in a simulation study for mixtures of Poisson data.