论文检索

可交换信念的调整AI 解读

We consider geometrically the practical, theoretical and computational treatment of second-order exchangeability within Bayes linear analysis. We establish Bayes linear sufficiency for sample means through the representation theorem for second-order exchangeable beliefs. Interrelationships between collections of Bayes linear belief adjustments are analysed through the resolution transforms induced by a second-order exchangeable sample. It is shown that the resolution…

SUMMARY We propose a general Bayesian method of comparing models. The approach is based on the Kullback-Leibler distance between two families of models, one nested within the other. For each parameter value of a full model, we compute the projection of the model to the restricted parameter space and the corresponding minimum distance. From the posterior distribution of the minimum distance, we can judge whether or not a more parsimonious model is appropriate. We show how the…

杂论：小波回归中的异方差得分检验AI 解读

We consider two score tests for heteroscedasticity in the errors of a signal-plus-noise model, where the signal is estimated by wavelet thresholding methods. The error variances are assumed to depend on observed covariates, through a parametric relationship of known form. The tests are based on the approaches of Breusch & Pagan (1979) and Koenker (1981). We establish the asymp totic validity of the tests and examine their performance in a simulation study. The Koenker test…

This paper presents a new methodological approach for carrying out Bayesian inference about dynamic models for exponential family observations. The approach is simulationbased and involves the use of Markov chain Monte Carlo techniques. A Metropolis- Hastings algorithm is combined with the Gibbs sampler in repeated use of an adjusted version of normal dynamic linear models. Different alternative schemes based on sampling from the system disturbances and state parameters…

序列匹配：分类时间序列的交叉谱分析AI 解读

We consider the problem of quantifying the degree to which two stationary categorical time series are coherent. The goal is to discover whether or not the sequences contain similar patterns. The problem is motivated by the problem of matching two DNA sequences. Following the ideas used in defining the spectral envelope for a qualitativevalued time series, the methods we present here focus on the problem of obtaining coherency envelopes for measuring the similarity between…

二元生存模型中关联结构的拟合优度检验AI 解读

We propose a simple test of constant conditional hazard ratio in the Clayton model (1978) by comparing the unweighted and weighted concordance estimators of the association parameter. If the Clayton model holds, the difference of these two estimates converges to zero. The proposed test is consistent against alternatives under which the two concordance estimators converge to different values. We derive an explicit formula for the asymptotic variance and derive the asymptotic…

比例优势回归与筛最大似然估计AI 解读

This paper introduces, under a censoring mechanism, an estimation procedure for the baseline function and the regression parameters in the proportional odds model based on the sieve maximum likelihood method. The procedure uses monotone splines of variable orders and knots as approximations for the baseline function and an approximately unbiased estimator of the comparative Kullback-Leibler risk to select the best approximation. The procedure adapts to various unknown…

竞争风险与失效原因掩蔽下的生存分析AI 解读

Consider a life testing situation in which systems are subject to failure from independent competing risks. The hazards of various risks are proportional to each other. When a failure occurs, immediate, i.e. stage 1, procedures are used in an attempt to reach a definitive diagnosis. If a diagnosis is not reached, this phenomenon is called masking. Stage 2 procedures, such as failure analysis or autopsy, provide definitive diagnosis for a small sample of the masked cases.…

具有依赖删失时间的不完整重复测量分析AI 解读

A frequently encountered complication in repeated-measurements data analysis is that a substantial number of subjects drop out of the study. In this paper, several conceptually simple methods are proposed for analysing incomplete repeated measures when the measurement times vary substantially from subject to subject and the subject's dropout time may depend on the observed and unobserved outcome variables. Numerical studies are conducted to examine the adequacy of our…

离散数据的半参数平滑AI 解读

A method for semiparametric smoothing of discrete data is proposed. The method consists of the repeated application of a Markov chain transition matrix constructed so as to have a given standard discrete parametric vehicle model as its stationary distribution. Theory and practical examples suggest that the approach yields improved performance over fully nonparametric methods when the vehicle model is a good one and otherwise provides a method comparable to fully…

正态线性模型的 Schwarz 准则及相关方法AI 解读

In this paper we derive Schwarz's information criterion and two modifications for choosing fixed effects in normal linear mixed models. The first modification allows an arbitrary, possibly informative, prior for the parameter of interest. Replacing this prior with the normal, unit-information, prior of Kass & Wasserman (1995) and the generalised Cauchy prior of Jeffreys (1961) yields the usual Schwarz criterion and a second modification, respectively. Under the null…

基于前三阶累积量的最优估计方程AI 解读

For improving on quasilikelihood estimation two types of quadratic estimating equations have been proposed, one based on the Edgeworth expansion, the other on the generalisation of the quasi-score. The first requires that the skewness of observations has a small departure from the exponential family; the second requires the knowledge of both skewness and kurtosis. We introduce an optimal quadratic estimating equation applicable when the skewness is not small and the kurtosis…

约束参数问题贝叶斯分析的蒙特卡洛方法AI 解读

Constraints on the parameters in a Bayesian hierarchical model typically make Bayesian computation and analysis complicated. Posterior densities that contain analytically intractable integrals as normalising constants depending on the hyperparameters often make implementation of Gibbs sampling or the Metropolis algorithms difficult. By using reweighting mixtures (Geyer, 1995), we develop alternative simulation-based methods to determine properties of the desired Bayesian…

从无偏线性估计方程到无偏估计量AI 解读

We suggest in this paper bias correction and prevention measures for the estimator defined as a solution to an unbiased linear estimating equation. A new formula for the exact bias of this estimator is presented and is expressed as a convergent sequence. Feasible exact bias correction based on the formula is computationally straightforward. Connections to Firth's (1993) device of bias prevention are drawn. The primary application is to the first-order autoregression model.

杂录：随机微分方程组的统计估计与模拟方法AI 解读

This paper proposes a new statistical method of estimation and simulation for systems of stochastic differential equations. By using this method, we can easily derive a discretised process from a system of stochastic differential equations. The discretised process is tractable enough to simulate the system and to construct the likelihood which is used to estimate the coefficients of the stochastic differential equations from discrete observation.

竞争风险模型中多种原因别风险率相等的一类检验AI 解读

A class of asymptotically distribution-free tests for the equality of the k risks in a competing risks model is considered. The tests can be derived by martingale theory or by a marginal likelihood approach. When the risks are dependent, the quantities to be compared would be the cause-specific hazard rates or equivalently the pseudosurvivor functions related to the cause-specific hazard rates. This class of tests allows one to compare k risks simultaneously for k > 2, is…

模型选择：最小后验预测损失方法AI 解读

Model choice is a fundamental and much discussed activity in the analysis of datasets. Nonnested hierarchical models introducing random effects may not be handled by classical methods. Bayesian approaches using predictive distributions can be used though the formal solution, which includes Bayes factors as a special case, can be criticised. We propose a predictive criterion where the goal is good prediction of a replicate of the observed data but tempered by fidelity to the…

杂记：聚类数据的病例对照抽样AI 解读

We derive the relationship between the population model and the sample model in the case of case-control sampling of clustered responses. The derived formulae provide insight into why methods that ignore the sampling scheme fail to provide consistent estimators in this case. The differences between clustered and independent responses in the context of the case-control design are discussed.

多元自回归移动平均模型的精确似然函数AI 解读

This paper derives the explicit expressions for the determinant and exact inverse of the covariance matrix of a multivariate autoregressive moving average model, and obtains the closed form of the exact likelihood function.

用于时间序列模型的扫描抽样器AI 解读

An algorithm, called the scan sampler, is developed and discussed. The scan sampler has a variety of uses for time series analysis based on the state space model with non-Gaussian observations. The algorithm is based on the Kalman filter/smoothing algorithm. It can be used for Bayesian inference using Markov chain Monte Carlo and to find posterior modes.

不完整数据下自回归移动平均模型的精确似然AI 解读

A transformation is developed which allows computationally efficient evaluation of the exact likelihood of an autoregressive-moving average model with incomplete data. Timings are presented which indicate that the new method is preferable to the Kalman filter in many instances. Extensions to nonstationary and vector models are discussed.

用于出版偏差建模的非参数权重函数类AI 解读

This paper addresses the use of weight functions to model publication bias in meta-analysis. Since publication bias is hard to gauge, a nonparametric ε-contamination class of weight functions is introduced. Sensitivity of conclusions to the specification of the weight function is explored by examining the range of results for the entire ε-contamination class. First, lower bounds are found on the coverage of confidence intervals. If little publication bias is suspected,…

从生命表到风险率：变换方法AI 解读

We address the problem of estimating hazard rates from lifetable data. The central issue is a discretisation bias, which is caused by the aggregation of lifetimes which occurs in lifetable data. Smoothing raw mortalities directly derived from lifetable data leads to severe biases whenever the true hazard rate is large. We propose here a new transformation of smoothed lifetable data which is superior to previous approaches. This is demonstrated by means of large-sample…

高斯独立增量群序贯设计中的偏差估计与减小AI 解读

A concern in the use of sequential testing procedures is the bias associated with the estimates of treatment differences. Clinical trials that stop early because of evidence of therapeutic benefit are prone to exaggerate the magnitude of the treatment effect. We consider methods for estimating and reducing the bias of treatment differences estimators in group sequential designs with Gaussian independent increment structure. We derive an analytical expression for the bias and…