基于广义最小二乘法三参数对数正态分布的渐近参数估计(英文)

通过二步迭代和广义 Gauss- Markov定理 ,给出了三参数对数正态分布参数的一种渐近估计 .这种方法既适合于全样本场合又适合于一般缺失数据场合 .模拟结果表明

半参数EV模型参数的二阶段估计

本文综合核函数法 ,最小二乘法 ,利用二阶段估计的方法求出了 EV模型中参数的估计量 ,并研究了它的强相合性以及渐近正态性 .

一类多维指数分布的参数估计

考虑生存函数为-F（x1,x2,……xn）=P{X1＞x1…,Xn＞xn}=exp{-[n∑i=1（xi/θi）♂]δ}（0＜xi＜∞,0＜δ≤1,0＜θi＜∞,i=-↑1,n 的一类多维指数分布，给出了它的密度函数的表示式，并讨论了它的性质．提出了相关参数δ的估计δ，证明了δ有相合性和渐近正态性，得到了δ的渐近方差δ．最后还给出了若干随机模拟的结果．

TWO-STEP ESTIMATORS IN PARTIAL LINEAR MODELS WITH MISSING RESPONSE VARIABLES AND ERROR-PRONE COVARIATES

A partial linear model with missing response variables and error-prone covariates is considered. The imputation approach is developed to estimate the regression coefficients and the nonparametric function. The proposed parametric estimators are shown to be asymptotically normal, and the estimators for the nonparametric part are proved to converge at an optimal rate. To construct confidence regions for the regression coefficients and the nonparametric function, respectively, the authors also propose the empirical-likelihood-based statistics and investigate the limit distributions of the empirical likelihood ratios. The simulation study is conducted to compare the finite sample behavior for the proposed estimators. An application to an AIDS dataset is illustrated.

Semi-parametric estimation for the Box-Cox transformation model with partially linear structure

The Box-Cox transformation model has been widely used in applied econometrics, positive accounting, positive finance and statistics. There is a large literature on Box-Cox transformation model with linear structure. However, there is seldom seen on the discussion for such a model with partially linear structure. Considering the importance of the partially linear model, in this paper, a relatively simple semi-parametric estimation procedure is proposed for the Box-Cox transformation model without presuming the linear functional form and without specifying any parametric form of the disturbance, which largely reduces the risk of model misspecification. We show that the proposed estimator is consistent and asymptotically normally distributed. Its covariance matrix is also in a closed form, which can be easily estimated. Finally, a simulation study is conducted to see the finite sample performance of our estimator.

Corrected empirical likelihood for a class of generalized linear measurement error models

Generalized linear measurement error models, such as Gaussian regression, Poisson regression and logistic regression, are considered. To eliminate the effects of measurement error on parameter estimation, a corrected empirical likelihood method is proposed to make statistical inference for a class of generalized linear measurement error models based on the moment identities of the corrected score function. The asymptotic distribution of the empirical log-likelihood ratio for the regression parameter is proved to be a Chi-squared distribution under some regularity conditions. The corresponding maximum empirical likelihood estimator of the regression parameter π is derived, and the asymptotic normality is shown. Furthermore, we consider the construction of the confidence intervals for one component of the regression parameter by using the partial profile empirical likelihood. Simulation studies are conducted to assess the finite sample performance. A real data set from the ACTG 175 study is used for illustrating the proposed method.

Asymptotic properties of Lasso in high-dimensional partially linear models

We study the properties of the Lasso in the high-dimensional partially linear model where the number of variables in the linear part can be greater than the sample size.We use truncated series expansion based on polynomial splines to approximate the nonparametric component in this model.Under a sparsity assumption on the regression coefficients of the linear component and some regularity conditions,we derive the oracle inequalities for the prediction risk and the estimation error.We also provide sufficient conditions under which the Lasso estimator is selection consistent for the variables in the linear part of the model.In addition,we derive the rate of convergence of the estimator of the nonparametric function.We conduct simulation studies to evaluate the finite sample performance of variable selection and nonparametric function estimation.

ON ASYMPTOTIC NORMALITY OF PARAMETERS IN MULTIPLE LINEAR ERRORS-IN-VARIABLES MODEL

This paper studies the parameter estimation of multiple dimensional linear errors-in-variables (EV) models in the case where replicated observations are available in some experimental points. Asymptotic normality is established under mild conditions, and the parameters entering the asymptotic variance are consistently estimated to render the result useable in the construction of large-sample confidence regions.

Asymptotically efficient parameter estimation for ordinary differential equations

Parameter estimation for ordinary differential equations arises in many fields of science and engineering. To be the best of our knowledge, traditional methods are often either computationally intensive or inaccurate for statistical inference. Ramsay et al.(2007) proposed a generalized profiling procedure. It is easily implementable and has been demonstrated to have encouraging numerical performance. However, little is known about statistical properties of this procedure. In this paper, we provide a theoretical justification of the generalized profiling procedure. Under some regularity conditions, the procedure is shown to be consistent for a broad range of tuning parameters. When the tuning parameters are sufficiently large, the procedure can be further shown to be asymptotically normal and efficient.