Amer Statist Research Articles

False Discovery Variance Reduction in Large Scale Simultaneous Hypothesis Tests

Statistical dependence between hypotheses poses a significant challenge to the stability of large scale multiple hypotheses testing. Ignoring it often results in an unacceptably large spread in the false positive proportion even though the average value is acceptable (Fan et al., J Amer Statist Assoc 107(499): 1019-1035, 2012; Owen J R Stat Soc Ser B 67(3): 411–426, 2005; Qiu et al., Stat Appl Genet Mol Biol 4: 32, 2005 and Schwartzman and Lin Biometrika 98(1): 199–214, 2011). However, the statistical dependence structure of data is often unknown. Using a generic signal-processing model, Bayesian multiple testing, and simulations, we demonstrate that the variance of the false positive proportion can be substantially reduced even under unknown short range dependence. We do this by modeling the data generating process as a stationary ergodic binary signal process embedded in noisy observations. We derive conditional probabilities needed for the Bayesian multiple testing by incorporating nearby observations into a second order Taylor series approximation. Simulations under general conditions are carried out to assess the validity and the variance reduction of the approach. Along the way, we address the problem of sampling a random Markov matrix with specified stationary distribution and lower bounds on the top absolute eigenvalues, which is of interest in its own right.

Strong consistency of kernel estimator in a semiparametric regression model

ABSTRACTEstimating the effective dimension reduction (EDR) space, related to the semiparametric regression model introduced by Li [Sliced inverse regression for dimension reduction. J Amer Statist Assoc. 1991;86:316–327], is based on the estimation of the covariance matrix Λ of the conditional expectation of the vector of predictors given the response. An estimator of Λ based on kernel method was introduced by Zhu and Fang [Asymptotics for kernel estimate of sliced inverse regression. Ann Statist. 1996;24:1053–1068] who then derived, under some conditions, the asymptotic distribution of , as . In this paper, we obtain the almost sure convergence of to Λ, as .

Pairwise directions estimation for multivariate response regression data

ABSTRACTThis article concerns the analysis of multivariate response data with multi-dimensional covariates. Based on local linear smoothing techniques, we propose an iteratively adaptive estimation method to reduce the dimensions of response variables and covariates. Two weighted estimation strategies are incorporated in our approach to provide initial estimates. Our proposal is also extended to curve response data for a data-adaptive basis function searching. Instead of focusing on goodness of fit, we shift the problem to reveal the data structure and basis patterns. Simulation studies with multivariate response and curve data are conducted for our pairwise directions estimation (PDE) approach in comparison with sliced inverse regression of Li et al. [Dimension reduction for multivariate response data. J Amer Statist Assoc. 2003;98:99–109]. The results demonstrate that the proposed PDE method is useful for data with responses approximating linear or bending structures. Illustrative applications to two real datasets are also presented.

Estimating coefficients of single-index regression models by minimizing variation

ABSTRACTThis paper proposes a novel estimation of coefficients in single-index regression models. Unlike the traditional average derivative estimation [Powell JL, Stock JH, Stoker TM. Semiparametric estimation of index coefficients. Econometrica. 1989;57(6):1403–1430; Hardle W, Thomas M. Investigating smooth multiple regression by the method of average derivatives. J Amer Statist Assoc. 1989;84(408):986–995] and semiparametric least squares estimation [Ichimura H. Semiparametric least squares (sls) and weighted sls estimation of single-index models. J Econometrics. 1993;58(1):71–120; Hardle W, Hall P, Ichimura H. Optimal smoothing in single-index models. Ann Statist. 1993;21(1):157–178], the procedure developed in this paper is to estimate the coefficients directly by minimizing the mean variation function and does not involve estimating the link function nonparametrically. As a result, it avoids the selection of the bandwidth or the number of knots, and its implementation is more robust and easier. The resultant estimator is shown to be consistent. Numerical results and real data analysis also show that the proposed procedure is more applicable against model free assumptions.

The Topp–Leone odd log-logistic family of distributions

ABSTRACTWe introduce a new class of continuous distributions named the Topp–Leone odd log-logistic family, which extends the one-parameter distribution pioneered by Topp and Leone [A family of J-shaped frequency functions. J Amer Statist Assoc. 1955;50:209–219]. We study some of its mathematical properties and describe two special cases. Further, we propose a regression model based on the new Topp–Leone odd log-logistic Weibull distribution. The usefulness and flexibility of the proposed family are illustrated by means of three real data sets.

Bayesian inference for Birnbaum–Saunders distribution and its generalization

ABSTRACTWe present a Bayesian approach for parameter inference of the Birnbaum–Saunders distribution [Birnbaum ZW, Saunders SC. A new family of life distributions. J Appl Probab. 1969;6:319–327], as well as the generalized Birnbaum–Saunders distribution developed by Owen [A new three-parameter extension to the Birnbaum–Saunders distribution. IEEE Trans Reliab. 2006;55:475–479], in the presence of random right-censored data. To handle the instance of commonly occurred censored observations, we utilize the data augmentation technique [Tanner MA, Wong WH. The calculation of posterior distributions by data augmentation. J Amer Statist Assoc. 1987;82(398):528–540] to circumvent the arduous expressions involving the censored data in posterior inferences. Simulation studies are carried out to assess performance of these methods under different parameter values, with small and large sample sizes, as well as various degrees of censoring. Two real data are analysed for illustrative purpose.

A generalized possibilistic approach to shell clustering of template-based shapes

ABSTRACTEfficiently and effectively detecting shell-like structures of particular shapes is an important task in computer vision and image processing. This paper presents a generalized possibilistic c-means algorithm (PCM) for shell clustering based on the diversity index of degree-λ proposed by Patil and Taillie [Diversity as a concept and its measurement. J Amer Statist Assoc. 1982;77:548–561]. Experiments on various data sets in Wang [Possibilistic shell clustering of template-based shapes. IEEE Trans Fuzzy Syst. 2009;17:777–793] show that the the proposed generalized PCM performs better than Wang's [Possibilistic shell clustering of template-based shapes. IEEE Trans Fuzzy Syst. 2009;17:777–793] possibilistic shell clustering method according two two criteria: (i) the ‘grade of detection’ for each target cluster; (ii) the amount of computation, denoted as , required to attain a given .

Doubly robust empirical likelihood inference in covariate-missing data problems

Missing covariate data occurs often in regression analysis. We study methods for estimating the regression coefficients in an assumed conditional mean function when some covariates are completely observed but other covariates are missing for some subjects. We adopt the semiparametric perspective of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Amer Statist Assoc. 1994;89:846–866] on regression analyses with missing covariates, in which they pioneered the use of two working models, the working propensity score model and the working conditional score model. A recent approach to missing covariate data analysis is the empirical likelihood method of Qin et al. [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503], which effectively combines unbiased estimating equations. In this paper, we consider an alternative likelihood approach based on the full likelihood of the observed data. This full likelihood-based method enables us to generate estimators for the vector of the regression coefficients that are (a) asymptotically equivalent to those of Qin et al. [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503] when the working propensity score model is correctly specified, and (b) doubly robust, like the augmented inverse probability weighting (AIPW) estimators of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Am Statist Assoc. 1994;89:846–866]. Thus, the proposed full likelihood-based estimators improve on the efficiency of the AIPW estimators when the working propensity score model is correct but the working conditional score model is possibly incorrect, and also improve on the empirical likelihood estimators of Qin, Zhang and Leung [Empirical likelihood in missing data problems. J Amer Statist Assoc. 2009;104:1492–1503] when the reverse is true, that is, the working conditional score model is correct but the working propensity score model is possibly incorrect. In addition, we consider a regression method for estimation of the regression coefficients when the working conditional score model is correctly specified; the asymptotic variance of the resulting estimator is no greater than the semiparametric variance bound characterized by the theory of Robins et al. [Estimation of regression coefficients when some regressors are not always observed. J Amer Statist Assoc. 1994;89:846–866]. Finally, we compare the finite-sample performance of various estimators in a simulation study.

Exact confidence intervals for the shape parameter of the gamma distribution

In this paper exact confidence intervals (CIs) for the shape parameter of the gamma distribution are constructed using the method of Bølviken and Skovlund [Confidence intervals from Monte Carlo tests. J Amer Statist Assoc. 1996;91:1071–1078]. The CIs which are based on the maximum likelihood estimator or the moment estimator are compared to bootstrap CIs via a simulation study.

Approximations to the p-values of tests for a change-point under non-standard conditions

Three test statistics for a change-point in a linear model, variants of those considered by Andrews and Ploberger [Optimal tests when a nusiance parameter is present only under the alternative. Econometrica. 1994;62:1383–1414]: the sup-likelihood ratio (LR) statistic; a weighted average of the exponential of LR-statistics and a weighted average of LR-statistics, are studied. Critical values for the statistics with time trend regressors, obtained via simulation, are found to vary considerably, depending on conditions on the error terms. The performance of the bootstrap in approximating p-values of the distributions is assessed in a simulation study. A sample approximation to asymptotic analytical expressions extending those of Kim and Siegmund [The likelihood ratio test for a change-point in simple linear regression. Biometrika. 1989;76:409–423] in the case of the sup-LR test is also assessed. The approximations and bootstrap are applied to the Quandt data [The estimation of a parameter of a linear regression system obeying two separate regimes. J Amer Statist Assoc. 1958;53:873–880] and real data concerning a change-point in oxygen uptake during incremental exercise testing and the bootstrap gives reasonable results.

Linear estimation for the extended exponential power distribution

Tiao and Lund [The use of OLUMV estimators in inference robustness studies of the location parameter of a class of symmetric distributions. J Amer Statist Assoc. 1970;65(329):370–386] tabulated the coefficients of the best linear unbiased estimators (BLUEs) of location and scale for a particular family of symmetric distributions. This family was a reparameterization of the extended exponential power distribution (EEPD) with the shape parameter restricted to be greater than or equal to one. In this work, we consider the BLU estimation of the location and scale parameters of the EEPD when the shape parameter is one-third and one-half. We obtain closed-form expressions for the single and product moments of the order statistics when the shape parameter is in general in the form of a reciprocal of an integer. These expressions are then used to determine the BLUEs and the corresponding variances for complete samples of size 20 and less. We consider some other linear estimators of the location and scale parameters and then compare them with the BLUEs. Finally, we present a numerical example to illustrate the developed results.

Simulating dependent binary variables through multinomial sampling

A new method for simulating dependent binary data is presented. This methodology does not require specification of a parametric joint distribution, and instead uses desired dependence levels and marginal probabilities to construct a cumulative distribution function for the multinomial distribution of all possible combinations of binary outcomes. This method is simple to use and compute, and through simulation studies is found to be comparable with the gold-standard approach by Emrich and Piedmonte [A method for generating high-dimensional multivariate binary variates. Amer Statist. 1991;45(4):302–304].

Quantile regression based on a weighted approach under semi-competing risks data

In this article, we investigate the quantile regression analysis for semi-competing risks data in which a non-terminal event may be dependently censored by a terminal event. Due to the dependent censoring, the estimation of quantile regression coefficients on the non-terminal event becomes difficult. In order to handle this problem, we assume Archimedean Copula to specify the dependence of the non-terminal event and the terminal event. Portnoy [Censored regression quantiles. J Amer Statist Assoc. 2003;98:1001–1012] considered the quantile regression model under right-censoring data. We extend his approach to construct a weight function, and then impose the weight function to estimate the quantile regression parameter for the non-terminal event under semi-competing risks data. We also prove the consistency and asymptotic properties for the proposed estimator. According to the simulation studies, the performance of our proposed method is good. We also apply our suggested approach to analyse a real data.

Estimation of cointegrated models with exogenous variables

We consider a cointegrated vector autoregressive process of integrated order 1, where the process consists of endogenous variables and exogenous variables. Johansen [Cointegration in partial systems and the efficiency of single-equation analysis. J Econometrics. 1992;52:389–402], Harbo et al. [Asymptotic inference on cointegrating rank in partial systems. J Amer Statist Assoc. 1998;16:388–399], and Pesaran et al. [Structural analysis of vector error correction models with exogenous I(1) variables. J Econometrics. 2000;97:293–343] considered inference of such processes assuming that the non-stationary exogenous variables are not cointegrated, and thus they are weakly exogenous. We consider the case where exogenous variables are cointegrated. Parameterization and estimation of the model is considered, and the asymptotic properties of the estimators are presented. The method in this paper is also applicable for the models considered in Mosconi and Giannini [Non-causality in cointegrated systems: representation estimation and testing. Oxford Bull Econ Stat. 1992;54:399–417], Pradel and Rault [Exogeneity in vector error correction models with purely exogenous long-run paths. Oxford Bull Econ Stat. 2003;65:629–653], and Hunter [Cointegrating exogeneity. Econom Lett. 1990;34:33–35]. A real data example is provided to illustrate the methods. Finite sample properties of the estimators are also examined through a Monte Carlo simulation.

A study of the probit model with latent variables in Phase I clinical trials

In recent years, the continual reassessment method (CRM) has gained considerable popularity in Phase I cancer studies. In this article, we propose a two-parameter probit model with latent variables. Introducing the latent variables enables exact Bayesian inference in the sense that conditional posterior distributions of all the parameters, including the latent variables, are known. Such inference can be more accurate than the maximum likelihood inference and can also alleviate the estimation issue due to small sample size [Albert J, Chib S. Bayesian analysis of binary and polychotomous response data. J Amer Statist Assoc. 1993;88:669–679] in Phase I studies. Simulation studies demonstrated that the proposed method compares favourably to the one-parameter CRM.

Variance function in regression analysis of longitudinal data using the generalized estimating equation approach

Longitudinal or clustered response data arise in many applications such as biostatistics, epidemiology and environmental studies. The repeated responses cannot in general be assumed to be independent. One method of analysing such data is by using the generalized estimating equations (GEE) approach. The current GEE method for estimating regression effects in longitudinal data focuses on the modelling of the working correlation matrix assuming a known variance function. However, correct choice of the correlation structure may not necessarily improve estimation efficiency for the regression parameters if the variance function is misspecified [Wang YG, Lin X. Effects of variance-function misspecification in analysis of longitudinal data. Biometrics. 2005;61:413–421]. In this connection two problems arise: finding a correct variance function and estimating the parameters of the chosen variance function. In this paper, we study the problem of estimating the parameters of the variance function assuming that the form of the variance function is known and then the effect of a misspecified variance function on the estimates of the regression parameters. We propose a GEE approach to estimate the parameters of the variance function. This estimation approach borrows the idea of Davidian and Carroll [Variance function estimation. J Amer Statist Assoc. 1987;82:1079–1091] by solving a nonlinear regression problem where residuals are regarded as the responses and the variance function is regarded as the regression function. A limited simulation study shows that the proposed method performs at least as well as the modified pseudo-likelihood approach developed by Wang and Zhao [A modified pseudolikelihood approach for analysis of longitudinal data. Biometrics. 2007;63:681–689]. Both these methods perform better than the GEE approach.

Statistical inference for a two unit warm standby system with dependent structure

Assuming that the joint distribution of the lifetimes of online unit and standby unit and the repair time of a failed unit in a two unit warm standby system is trivariate exponential, measures of system performance such as system reliability, MTBF, point availability and steady state availability are obtained. see Marshall and Olkin (J Amer Statist Assoc 62:30–44, 1967). Further, a 100(1-α) % asymptotic confidence interval for steady state availability of the system and an estimator of system reliability based on moments are obtained. Numerical work is carried out to illustrate the behaviour of steady state availability as well as the system reliability based on moments by simulating samples from trivariate exponential distribution.

Moments of Rand's C statistic in cluster analysis

Two different formulations on the moments of Rand's C appear incompatible. The mean and variance of Rand's C statistic given by Fowlkes and Mallows (J. Amer Statist. Assoc. 78 (1983) 553) are special cases of the mean and variance of C given by DuBien and Warde (ASA proceedings of the social statistics section (1981) 309).

Volume, Surface Area, and the Harmonic Mean

1. E. Barbeau, Perron's result and a decision on admission tests, this MAGAZINE 59: 1 (1986), 12-22. 2. D. Harville, The use of linear-model methodology to rate high school or college football teams, J. Amer Statist. Assoc. 72 (1977), 278. 3. T. A. Jech, The ranking of incomplete tournaments: a mathematician's guide to popular sports, Amer. Math. Monthly 90 (1983), 246-266. 4. J. P. Keener, The Perron-Frobenius theorem and the ranking of football teams, SIAM Review 35:1 (1993), 80-93.