关于分层线性模型样本容量问题的研究

文章运用lackknife和Bootstrap的方法,对参数估计的方差进行改进,构造了合适的参数估计的置信区间.通过样本组数和组内个体数的变化,利用数据模拟的方法进行研究,表明参数估计的可靠性很大程度上依赖于组数;对于固定效应参数,组数取30就可以得到可靠的估计值.对于σ和方差协方差成分T,组数分别取50和70才能得到可靠的估计.     

分层线性模型中的经验贝叶斯与完全贝叶斯方法及其比较

文章对基于最大似然估计(ML)和基于约束最大似然估计(REML)的经验贝叶斯方法(EB)进行了比较;指出了经验贝叶斯和完全贝叶斯方法各自存在的问题,并对两种方法进行了比较:给出了关于应用中如何选择推断方法的建议.     

基于分层广义线性模型的非寿险费率厘定精算模型研究

非寿险业务中的损失数据结构日益复杂,呈现异质性与相关性并存的异象。分层广义线性模型能够突破传统费率厘定精算方法仅分析风险个体同一保单年损失数据的局限,可以提高复杂结构损失数据预测的准确性。基于分层广义线性模型等方法,研究具有多年损失数据的非寿险费率厘定问题,并以车险和工伤补偿保险的两组损失数据为例进行实证分析。研究结果表明,相对于GLM而言,考虑随机效应后GLMM的拟合优度大幅改善,GLMM与HGLM可以更有效地反映不同风险个体的差异,并有利于揭示风险个体在多个保险期内损失的异质性与相关性。     

线性模型中的测量误差问题与工具变量方法

文章对简单线性回归模型中解释变量的测量误差的含义、后果以及相应的数理过程和统计性质作了较详细的说明和阐释,并就处理这类问题的工具变量方法的作用从理论和应用上进行了细致的分析。     

省域个人教育收益率变因剖析——基于分层线性模型视角

利用中国社会综合调查开放数据库（CGSS）中《中国城乡居民生活综合调查》（2005年）的数据资料和分层线性回归技术,对中国省域教育收益率及其影响因素进行实证分析。结果表明：现阶段,中国教育收益率已经高于世界平均水平;中国城乡居民男性教育收益率高于女性;少数民族教育收益率高于汉族;教育收益率随着受教育程度的提高而下降,但随着收入水平的提高而上升。     

分层线性模型的最大后验估计

最大后验估计（MAPE）和最大似然估计（MLE）都是重要的参数点估计方法。在介绍一般分层线性模型（HLM）MAPE方法的基础上,给出这种方法的期望最大化算法（EM）的具体步骤,运用对数似然函数的二阶导数推导了MAPE估计的方差估计量。同时运用数据模拟比较了EM算法下的MAPE和MLE。对于固定效应的估计,两种方法得到的估计量是一致的。当组数较少时,EM计算的MAPE的方差协方差成分比MLE的更靠近真实值,而且MAPE的迭代次数明显小于MLE。     

联合广义线性模型中的变量选择

在联合广义线性模型中,散度参数与均值都被赋予了广义线性模型的结构,本文主要考虑在只有分布的一阶矩和二阶矩指定的条件下,联合广义线性模型中均值部分的变量选择问题。本文采用广义拟似然函数,提出了新的模型选择准则(EAIC);该准则是Akaike信息准则的推广。论文通过模拟研究验证了该准则的效果。     

多层线性模型的参数估计

文章详对多层线性模型中固定效应与随机效应参数估计方法进行了理论推导,并对推导出的估计量进行了详细说明。     

单变量Bayes正态动态线性模型及其预测

Bayes预测和动态模型是20世纪70年代发展起来的一套新的时间序列分析方法,其中单变量Bayes正态动态线性模型(UBNDLM)在实际应用中最为常见和重要。在UBNDLM应用中,通常假定观测误差方差是未知常量,其值在建模的开始由估计给出。一般的做法是对具体的问题通过专家经验给出,并没有一个统一有效的办法。文章针对这一情形,给出了观测误差方差值一种简单易用的估计方法。同时也提供了数值试验说明新方法是有效的,这就给该类UBNDLM的应用带来了方便。     

基于多层线性模型的就业影响因素研究

当前就业问题已成为困扰各国政府的世界性难题,对于我国这样一个人口大国尤为突出。文章针对我国就业问题,建立了影响我国就业情况的分层模式指标体系,且通过以就业人数为因变量,各影响因素为自变量的逐步回归模型消除影响因素间多重共线性,得到了影响就业的主要因素。在此基础上,以样本单位为第一水平,样本所在地区为第二水平,建立了关于就业的分地区二层线性模型。     

Hierarchical CUB Models for Ordinal Variables

Hierarchical CUB models are a generalization of CUB models in which parameters are allowed to be random. The main feature that distinguishes such proposal from the standard one is the modeling of variation among groups. We illustrate the usefulness of these hierarchical structures by discussing model specification, inferential issues, and empirical results with reference to a real data set.     

Estimation for Partially Linear Single-index Instrumental Variables Models

In this article, we generalize the partially linear single-index models to the scenario with some endogenous covariates variables. It is well known that the estimators based on the existing methods are often inconsistent because of the endogeneity of covariates. To deal with the endogenous variables, we introduce some auxiliary instrumental variables. A three-stage estimation procedure is proposed for partially linear single-index instrumental variables models. The first stage is to obtain a linear projection of endogenous variables on a set of instrumental variables, the second stage is to estimate the link function by using local linear smoother for given constant parameters, and the last stage is to obtain the estimators of constant parameters based on the estimating equation. Asymptotic normality is established for the proposed estimators. Some simulation studies are undertaken to assess the finite sample performance of the proposed estimation procedure.     

Double Sampling Designs in Multivariate Linear Models with Missing Variables

We provide a method for finding the optimal double sampling plan for estimating the mean value of a continuous outcome. It is assumed that the fallible and true outcome data are related by a multivariate linear regression model where only some of the explanatory variables are sampled. Conditions under which double sampling is preferred over standard sampling plans are determined. An application of the method to a well-known data set on air pollution is presented.     

Hierarchical Generalized Linear Models and Frailty Models with Bayesian Nonparametric Mixing

This paper proposes Bayesian nonparametric mixing for some well-known and popular models. The distribution of the observations is assumed to contain an unknown mixed effects term which includes a fixed effects term, a function of the observed covariates, and an additive or multiplicative random effects term. Typically these random effects are assumed to be independent of the observed covariates and independent and identically distributed from a distribution from some known parametric family. This assumption may be suspect if either there is interaction between observed covariates and unobserved covariates or the fixed effects predictor of observed covariates is misspecified. Another cause for concern might be simply that the covariates affect more than just the location of the mixed effects distribution. As a consequence the distribution of the random effects could be highly irregular in modality and skewness leaving parametric families unable to model the distribution adequately. This paper therefore proposes a Bayesian nonparametric prior for the random effects to capture possible deviances in modality and skewness and to explore the observed covariates' effect on the distribution of the mixed effects.     

Generalized Linear Latent Variable Models with Flexible Distribution of Latent Variables

Abstract. We consider a semi‐nonparametric specification for the density of latent variables in Generalized Linear Latent Variable Models (GLLVM). This specification is flexible enough to allow for an asymmetric, multi‐modal, heavy or light tailed smooth density. The degree of flexibility required by many applications of GLLVM can be achieved through this semi‐nonparametric specification with a finite number of parameters estimated by maximum likelihood. Even with this additional flexibility, we obtain an explicit expression of the likelihood for conditionally normal manifest variables. We show by simulations that the estimated density of latent variables capture the true one with good degree of accuracy and is easy to visualize. By analysing two real data sets we show that a flexible distribution of latent variables is a useful tool for exploring the adequacy of the GLLVM in practice.     

An Automated MCEM Algorithm for Hierarchical Models with Multivariate and Multitype Response Variables

In this article, we consider a model allowing the analysis of multivariate data, which can contain data attributes of different types (e.g., continuous, discrete, binary). This model is a two-level hierarchical model which supports a wide range of correlation structures and can accommodate overdispersed data. Maximum likelihood estimation of the model parameters is achieved with an automated Monte Carlo expectation maximization algorithm. Our method is tested in a simulation study in the bivariate case and applied to a data set dealing with beehive activity.     

Modified Normal-based Approximation to the Percentiles of Linear Combination of Independent Random Variables with Applications

A modified normal-based approximation for calculating the percentiles of a linear combination of independent random variables is proposed. This approximation is applicable in situations where expectations and percentiles of the individual random variables can be readily obtained. The merits of the approximation are evaluated for the chi-square and beta distributions using Monte Carlo simulation. An approximation to the percentiles of the ratio of two independent random variables is also given. Solutions based on the approximations are given for some classical problems such as interval estimation of the normal coefficient of variation, survival probability, the difference between or the ratio of two binomial proportions, and for some other problems. Furthermore, approximation to the percentiles of a doubly noncentral F distribution is also given. For all the problems considered, the approximation provides simple satisfactory solutions. Two examples are given to show applications of the approximation.     

Empirical Likelihood for Partially Non Linear Models with Missing Response Variables at Random

This article is concerned with partially non linear models when the response variables are missing at random. We examine the empirical likelihood (EL) ratio statistics for unknown parameter in non linear function based on complete-case data, semiparametric regression imputation, and bias-corrected imputation. All the proposed statistics are proven to be asymptotically chi-square distribution under some suitable conditions. Simulation experiments are conducted to compare the finite sample behaviors of the proposed approaches in terms of confidence intervals. It showed that the EL method has advantage compared to the conventional method, and moreover, the imputation technique performs better than the complete-case data.     

MCMC Sampler Convergence Rates for Hierarchical Normal Linear Models: A Simulation Approach

This paper presents a straightforward method of approximating theoretical bounds on burn-in time for MCMC samplers for hierarchical normal linear models. An extension and refinement of Cowles and Rosenthal's (1998) simulation approach, it exploits Hodges's (1998) reformulation of hierarchical normal linear models. The method is illustrated with three real datasets, involving a one-way variance components model, a growth-curve model, and a spatial model with a pairwise-differences prior. In all three cases, when the specified priors produce proper, unimodal posterior distributions, the method provides very reasonable upper bounds on burn-in time. In contrast, when the posterior distribution for the variance-components model can be shown to be improper or bimodal, the new method correctly identifies convergence failure while several other commonly-used diagnostics provide false assurance that convergence has occurred.