In this paper, we propose a new framework to construct confidence sets for a $d$-dimensional unknown sparse parameter ${\boldsymbol \theta }$ under the normal mean model ${\boldsymbol X}\sim N({\boldsymbol \theta },\sigma ^{2}\bf{I})$. A key feature of the proposed confidence set is its capability to account for the sparsity of ${\boldsymbol \theta }$, thus named as sparse confidence set. This is in sharp contrast with the classical methods, such as the Bonferroni confidence intervals and other resampling-based procedures, where the sparsity of ${\boldsymbol \theta }$ is often ignored. Specifically, we require the desired sparse confidence set to satisfy the following two conditions: (i) uniformly over the parameter space, the coverage probability for ${\boldsymbol \theta }$ is above a pre-specified level; (ii) there exists a random subset $S$ of $\{1,...,d\}$ such that $S$ guarantees the pre-specified true negative rate for detecting non-zero $\theta _{j}$’s. To exploit the sparsity of ${\boldsymbol \theta }$, we allow the confidence interval for $\theta _{j}$ to degenerate to a single point 0 for any $j\notin S$. Under this new framework, we first consider whether there exist sparse confidence sets that satisfy the above two conditions. To address this question, we establish a non-asymptotic minimax lower bound for the non-coverage probability over a suitable class of sparse confidence sets. The lower bound deciphers the role of sparsity and minimum signal-to-noise ratio (SNR) in the construction of sparse confidence sets. Furthermore, under suitable conditions on the SNR, a two-stage procedure is proposed to construct a sparse confidence set. To evaluate the optimality, the proposed sparse confidence set is shown to attain a minimax lower bound of some properly defined risk function up to a constant factor. Finally, we develop an adaptive procedure to the unknown sparsity. Numerical studies are conducted to verify the theoretical results.
Varying coefficient models have been used to explore dynamic effects in many scientific areas, such as in medicine, finance, and epidemiology. As most existing models ignore the existence of zero regions, we propose a new soft-thresholded varying coefficient model, where the coefficient functions are piecewise smooth with zero regions. Our new modeling approach enables us to perform variable selection, detect the zero regions of selected variables, obtain point estimates of the varying coefficients with zero regions, and construct a new type of sparse confidence intervals that accommodate zero regions. We prove the asymptotic properties of the estimator, based on which we draw statistical inference. Our simulation study reveals that the proposed sparse confidence intervals achieve the desired coverage probability. We apply the proposed method to analyze a large-scale preoperative opioid study.
more » « less- PAR ID:
- 10432533
- Publisher / Repository:
- Oxford University Press
- Date Published:
- Journal Name:
- Biometrics
- Volume:
- 79
- Issue:
- 4
- ISSN:
- 0006-341X
- Format(s):
- Medium: X Size: p. 3388-3401
- Size(s):
- p. 3388-3401
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
more » « less
Abstract For statistical inference on regression models with a diverging number of covariates, the existing literature typically makes sparsity assumptions on the inverse of the Fisher information matrix. Such assumptions, however, are often violated under Cox proportion hazards models, leading to biased estimates with under‐coverage confidence intervals. We propose a modified debiased lasso method, which solves a series of quadratic programming problems to approximate the inverse information matrix without posing sparse matrix assumptions. We establish asymptotic results for the estimated regression coefficients when the dimension of covariates diverges with the sample size. As demonstrated by extensive simulations, our proposed method provides consistent estimates and confidence intervals with nominal coverage probabilities. The utility of the method is further demonstrated by assessing the effects of genetic markers on patients' overall survival with the Boston Lung Cancer Survival Cohort, a large‐scale epidemiology study investigating mechanisms underlying the lung cancer.
-
more » « less
Abstract Varying coefficient models are a flexible extension of generic parametric models whose coefficients are functions of a set of effect-modifying covariates instead of fitted constants. They are capable of achieving higher model complexity while preserving the structure of the underlying parametric models, hence generating interpretable predictions. In this paper we study the use of gradient boosted decision trees as those coefficient-deciding functions in varying coefficient models with linearly structured outputs. In contrast to the traditional choices of splines or kernel smoothers, boosted trees are more flexible since they require no structural assumptions in the effect modifier space. We introduce our proposed method from the perspective of a localized version of gradient descent, prove its theoretical consistency under mild assumptions commonly adapted by decision tree research, and empirically demonstrate that the proposed tree boosted varying coefficient models achieve high performance qualified by their training speed, prediction accuracy and intelligibility as compared to several benchmark algorithms.
-
null (Ed.)Abstract Zero-inflated and hurdle models are widely applied to count data possessing excess zeros, where they can simultaneously model the process from how the zeros were generated and potentially help mitigate the effects of overdispersion relative to the assumed count distribution. Which model to use depends on how the zeros are generated: zero-inflated models add an additional probability mass on zero, while hurdle models are two-part models comprised of a degenerate distribution for the zeros and a zero-truncated distribution. Developing confidence intervals for such models is challenging since no closed-form function is available to calculate the mean. In this study, generalized fiducial inference is used to construct confidence intervals for the means of zero-inflated Poisson and Poisson hurdle models. The proposed methods are assessed by an intensive simulation study. An illustrative example demonstrates the inference methods.more » « less
-
more » « less
Summary Benjamini and Yekutieli suggested that it is important to account for multiplicity correction for confidence intervals when only some of the selected intervals are reported. They introduced the concept of the false coverage rate (FCR) for confidence intervals which is parallel to the concept of the false discovery rate in the multiple-hypothesis testing problem and they developed confidence intervals for selected parameters which control the FCR. Their approach requires the FCR to be controlled in the frequentist’s sense, i.e. controlled for all the possible unknown parameters. In modern applications, the number of parameters could be large, as large as tens of thousands or even more, as in microarray experiments. We propose a less conservative criterion, the Bayes FCR, and study confidence intervals controlling it for a class of distributions. The Bayes FCR refers to the average FCR with respect to a distribution of parameters. Under such a criterion, we propose some confidence intervals, which, by some analytic and numerical calculations, are demonstrated to have the Bayes FCR controlled at level q for a class of prior distributions, including mixtures of normal distributions and zero, where the mixing probability is unknown. The confidence intervals are shrinkage-type procedures which are more efficient for the θis that have a sparsity structure, which is a common feature of microarray data. More importantly, the centre of the proposed shrinkage intervals reduces much of the bias due to selection. Consequently, the proposed empirical Bayes intervals are always shorter in average length than the intervals of Benjamini and Yekutieli and can be only 50% or 60% as long in some cases. We apply these procedures to the data of Choe and colleagues and obtain similar results.
