Bootstrap adaptive estimation: The trimmed-mean example

We consider the problem of choosing among a class of possible estimators by selecting the estimator with the smallest bootstrap estimate of finite sample variance. This is an alternative to using cross-validation to choose an estimator adaptively. The problem of a confidence interval based on such an adaptive estimator is considered. We illustrate the ideas by applying the method to the problem of choosing the trimming proportion of an adaptive trimmed mean. It is shown that a bootstrap adaptive trimmed mean is asymptotically normal with an asymptotic variance equal to the smallest among trimmed means. The asymptotic coverage probability of a bootstrap confidence interval based on such adaptive estimators is shown to have the nominal level. The intervals based on the asymptotic normality of the estimator share the same asymptotic result, but have poor small-sample properties compared to the bootstrap intervals. A small-sample simulation demonstrates that bootstrap adaptive trimmed means adapt themselves rather well even for samples of size 10.     

AN ADAPTIVE TRIMMED LIKELIHOOD ALGORITHM FOR IDENTIFICATION OF MULTIVARIATE OUTLIERS

This article describes an algorithm for the identification of outliers in multivariate data based on the asymptotic theory for location estimation as described typically for the trimmed likelihood estimator and in particular for the minimum covariance determinant estimator. The strategy is to choose a subset of the data which minimizes an appropriate measure of the asymptotic variance of the multivariate location estimator. Observations not belonging to this subset are considered potential outliers which should be trimmed. For α less than about 0.5, the correct trimming proportion is taken to be that α > 0 for which the minimum of any minima of this measure of the asymptotic variance occurs. If no minima occur for an α > 0 then the data set will be considered outlier free.     

Trimmed least squares estimator as best trimmed linear conditional estimator for linear regression model

A class of trimmed linear conditional estimators based on regression quantiles for the linear regression model is introduced. This class serves as a robust analogue of non-robust linear unbiased estimators. Asymptotic analysis then shows that the trimmed least squares estimator based on regression quantiles ( Koenker and Bassett ( 1978 ) ) is the best in this estimator class in terms of asymptotic covariance matrices. The class of trimmed linear conditional estimators contains the Mallows-type bounded influence trimmed means ( see De Jongh et al ( 1988 ) ) and trimmed instrumental variables estimators. A large sample methodology based on trimmed instrumental variables estimator for confidence ellipsoids and hypothesis testing is also provided.     

A Note on Robustness of the β-Trimmed Mean

Qualitative robustness of the β-trimmed mean has already been observed in terms of relative efficiency and weak continuity of that estimator in neighbourhoods of the exponential distribution. Two more robustness considerations are given here in favour of the β-trimmed mean: the statistical functional representing this estimator is Fréchet differentiable; and it is a special case of the trimmed likelihood estimator. Further, simulations suggest that a fixed proportion of trimming is preferable to adaptive estimation in this case.     

An evaluation of the trimmed mean approach in clinical trials with dropout

The trimmed mean is a method of dealing with patient dropout in clinical trials that considers early discontinuation of treatment a bad outcome rather than leading to missing data. The present investigation is the first comprehensive assessment of the approach across a broad set of simulated clinical trial scenarios. In the trimmed mean approach, all patients who discontinue treatment prior to the primary endpoint are excluded from analysis by trimming an equal percentage of bad outcomes from each treatment arm. The untrimmed values are used to calculated means or mean changes. An explicit intent of trimming is to favor the group with lower dropout because having more completers is a beneficial effect of the drug, or conversely, higher dropout is a bad effect. In the simulation study, difference between treatments estimated from trimmed means was greater than the corresponding effects estimated from untrimmed means when dropout favored the experimental group, and vice versa. The trimmed mean estimates a unique estimand. Therefore, comparisons with other methods are difficult to interpret and the utility of the trimmed mean hinges on the reasonableness of its assumptions: dropout is an equally bad outcome in all patients, and adherence decisions in the trial are sufficiently similar to clinical practice in order to generalize the results. Trimming might be applicable to other inter‐current events such as switching to or adding rescue medicine. Given the well‐known biases in some methods that estimate effectiveness, such as baseline observation carried forward and non‐responder imputation, the trimmed mean may be a useful alternative when its assumptions are justifiable.     

Two-Stage Welsh's Trimmed Mean for the Simultaneous Equations Model

This paper discusses the large sample theory of the two-stage Welsh's trimmed mean for the limited information simultaneous equations model. Besides having asymptotic normality, this trimmed mean, as the two-stage least squares estimator, is a generalized least squares estimator. It also acts as a robust Aitken estimator for the simultaneous equations model. Examples illustrate real data analysis and large sample inferences based on this trimmed mean.     

Monte Carlo studies of some adaptive robust procedures for location

One linear and two nonlinear adaptive robust procedures have been developed in which preliminary statistics, based on tail lengths, attempt to identify distributions from which the samples arise so that a suitable robust estimator based on trimmed means can be used to estimate the location parameter. The efficiencies of the estimators based on the three proposed adaptive robust procedures have been obtained using Monte Carlo methods involving eight distributions and these efficiencies are compared with the efficiencies of nineteen other robust estimators.     

Comparison of estimation methods for the finite population mean in simple random sampling: symmetric super-populations

In this paper, a new estimator combined estimator (CE) is proposed for estimating the finite population mean ¯ Y _N in simple random sampling assuming a long-tailed symmetric super-population model. The efficiency and robustness properties of the CE is compared with the widely used and well-known estimators of the finite population mean ¯ Y _N by Monte Carlo simulation. The parameter estimators considered in this study are the classical least squares estimator, trimmed mean, winsorized mean, trimmed L-mean, modified maximum-likelihood estimator, Huber estimator (W24) and the non-parametric Hodges–Lehmann estimator. The mean square error criteria are used to compare the performance of the estimators. We show that the CE is overall more efficient than the other estimators. The CE is also shown to be more robust for estimating the finite population mean ¯ Y _N , since it is insensitive to outliers and to misspecification of the distribution. We give a real life example.     

A monte carlo comparison of regression trimmed means

Two proposals are made for constructing adaptive estimators of the parameters in a linear regression model. These estimators are based on regression trimmed means and use an idea of Jaeckel [(1971) Ann Math Statist 42, 1540-1552] and the bootstrap respectively. These adaptive trimmed means as well as some nonadaptive trimmed means are studied by Monte Carlo. A one-step biweight is also included for comparison purposes.     

Asymptotics for an Adaptive Trimmed Likelihood Location Estimator

An asymptotic normality result is given for an adaptive trimmed likelihood estimator of location, which parallels the asymptotic normality result for the adaptive trimmed mean. The new result comes out of studying the adaptive trimmed likelihood estimator modelled parametrically by a normal family but then examining the behavior when the underlying distribution is in fact some F different from normal. The asymptotic variance of the adaptive estimator is equal to the asymptotic variance of the trimmed likelihood estimator at the optimal trimming proportion for the distribution F, subject to that trimming proportion being positive and F being suitably smooth.     

On the Tukey depth of an atomic measure

This paper gives a relation between the convex Tukey trimmed region (see [J.C. Massé, R. Theodorescu, Halfplane trimming for bivariate distributions, J. Multivariate Anal. 48(2) (1994) 188–202]) of an atomic measure and the support of the measure. It is shown that an atomic measure is concentrated on the extreme points of its Tukey trimmed region. A property concerning the extreme points which have 0 mass is given. As a corollary, we give a new method of proof of the Koshevoy characterization result.     

Simultaneous confidence intervals in the analysis of orthogonal saturated designs

We present the first known method of constructing exact simultaneous confidence intervals for the analysis of orthogonal, saturated factorial designs. Given m independent, normally distributed, unbiased estimators of treatment contrasts, if there is an independent chi-squared estimator of error variance, then simultaneous confidence intervals based on the Studentized maximum modulus distribution are exact under all parameter configurations. In this paper, an analogous method is developed for the case of an orthogonal saturated design, for which the treatment contrasts are independently estimable but there is no independent estimator of error variance. Lacking an independent estimator of the error variance, the smallest sums of squares of effect estimators are pooled. The simultaneous confidence intervals are based on a probability inequality, for which the simultaneous confidence coefficient is achieved in the null case.     

Test inversion bootstrap confidence intervals

In this paper we explore the theoretical and practical implications of using bootstrap test inversion to construct confidence intervals. In the presence of nuisance parameters, we show that the coverage error of such intervals is O ( n ^−1/2) which may be reduced to O ( n ⁻¹) if a Studentized statistic is used. We present three simulation studies and compare the performance of test inversion methods with established methods on the problem of estimating a confidence interval for the dose–response parameter in models of the Japanese atomic bomb survivors data.     

Computing least trimmed squares regression with the forward search

Least trimmed squares (LTS) provides a parametric family of high breakdown estimators in regression with better asymptotic properties than least median of squares (LMS) estimators. We adapt the forward search algorithm of Atkinson (1994) to LTS and provide methods for determining the amount of data to be trimmed. We examine the efficiency of different trimming proportions by simulation and demonstrate the increasing efficiency of parameter estimation as larger proportions of data are fitted using the LTS criterion. Some standard data examples are analysed. One shows that LTS provides more stable solutions than LMS.     

Investigation of the performance of trimmed estimators of life time distributions with censoring

For the lifetime (or negative) exponential distribution, the trimmed likelihood estimator has been shown to be explicit in the form of a β‐trimmed mean which is representable as an estimating functional that is both weakly continuous and Fréchet differentiable and hence qualitatively robust at the parametric model. It also has high efficiency at the model. The robustness is in contrast to the maximum likelihood estimator (MLE) involving the usual mean which is not robust to contamination in the upper tail of the distribution. When there is known right censoring, it may be perceived that the MLE which is the most asymptotically efficient estimator may be protected from the effects of ‘outliers’ due to censoring. We demonstrate that this is not the case generally, and in fact, based on the functional form of the estimators, suggest a hybrid defined estimator that incorporates the best features of both the MLE and the β‐trimmed mean. Additionally, we study the pure trimmed likelihood estimator for censored data and show that it can be easily calculated and that the censored observations are not always trimmed. The different trimmed estimators are compared by a modest simulation study.     

Robust ridge and robust Liu estimator for regression based on the LTS estimator

In the multiple linear regression analysis, the ridge regression estimator and the Liu estimator are often used to address multicollinearity. Besides multicollinearity, outliers are also a problem in the multiple linear regression analysis. We propose new biased estimators based on the least trimmed squares (LTS) ridge estimator and the LTS Liu estimator in the case of the presence of both outliers and multicollinearity. For this purpose, a simulation study is conducted in order to see the difference between the robust ridge estimator and the robust Liu estimator in terms of their effectiveness; the mean square error. In our simulations, the behavior of the new biased estimators is examined for types of outliers: X-space outlier, Y-space outlier, and X-and Y-space outlier. The results for a number of different illustrative cases are presented. This paper also provides the results for the robust ridge regression and robust Liu estimators based on a real-life data set combining the problem of multicollinearity and outliers.     

Cochran's rule for simple random sampling

Cochran's rule for the minimum sample size to ensure adequate coverage of nominal 95% confidence intervals is derived by using the Edgeworth expansion for the distribution function of the standardized sample mean. The rule is extended for confidence intervals based on the Studentized sample mean. The performance of the rule and Edgeworth approximations for smaller sample sizes are examined by simulation.     

A comparison of robust estimators based on two types of trimming

The least trimmed squares (LTS) estimator and the trimmed mean (TM) are two well-known trimming-based estimators of the location parameter. Both estimates are used in practice, and they are implemented in standard statistical software (e.g., S-PLUS, R, Matlab, SAS). The breakdown point of each of these estimators increases as the trimming proportion increases, while the efficiency decreases. Here we have shown that for a wide range of distributions with exponential and polynomial tails, TM is asymptotically more efficient than LTS as an estimator of the location parameter, when they have equal breakdown points.     

Linear trimmed means for the linear regression with AR(1) errors model

For the linear regression with AR(1) errors model, the robust generalized and feasible generalized estimators of Lai et al. (2003) of regression parameters are shown to have the desired property of a robust Gauss Markov theorem. This is done by showing that these two estimators are the best among classes of linear trimmed means. Monte Carlo and data analysis for this technique have been performed.     

Feasible robust estimator in restricted semiparametric regression models based on the LTS approach

Under some nonstochastic linear restrictions based on either additional information or prior knowledge in a semiparametric regression model, a family of feasible generalized robust estimators for the regression parameter is proposed. The least trimmed squares (LTS) method proposed by Rousseeuw as a highly robust regression estimator is a statistical technique for fitting a regression model based on the subset of h observations (out of n) whose least-square fit possesses the smallest sum of squared residuals. The coverage h may be set between n/2 and n. The LTS estimator involves computing the hyperplane that minimizes the sum of the smallest h squared residuals. For practical purpose, it is assumed that the covariance matrix of the error term is unknown and thus feasible estimators are replaced. Then, we develop an algorithm for the LTS estimator based on feasible methods. Through the Monte Carlo simulation studies and a real data example, performance of the feasible type of robust estimators is compared with the classical ones in restricted semiparametric regression models.