异方差模型参数估计的有效性研究

本文运用随机模拟方法,对误差序列异方差模型中加权最小二乘（GLS）估计的有效性进行研究。研究表明,GLS估计的有效性与异方差强度有关,当异方差强度较强时,GLS估计比普通最小二乘（OLS）估计有效;当异方差强度较弱时,GLS估计不如OLS估计有效。     

The Wheels of the Orthogonal Latin Squares Polytope: Classification and Valid Inequalities

A wheel in a graph G(V,E) is an induced subgraph consisting of an odd hole and an additional node connected to all nodes of the hole. In this paper, we study the wheels of the intersection graph of the Orthogonal Latin Squares polytope (P_I). Our work builds on structural properties of wheels which are used to categorise them into a number of collectively exhaustive classes. These classes give rise to families of inequalities that are valid for P_I and facet-defining for its set-packing relaxation. The classification introduced allows us to establish the cardinality of the whole wheel class and determine the range of the coefficients of any variable included in a lifted wheel inequality. Finally, based on this classification, a constant-time recognition algorithm for wheel-inducing circulant matrices, is introduced.     

辨识被测凸轮轴安装偏心量的对径谐波法

凸轮轴在检测仪上的安装偏心,影响凸轮轴上凸轮廓形尺寸的测量精度。本文提出一种辨识被测凸轮轴安装偏心量的对径谐波法,证明了用对径谐波法所辨识的偏心量符合最小二乘准则。此法在实际测量中获得了满意的效果     

Basic structure of the asymptotic theory in dynamic nonlineaerco nometric models, part i: consistency and approximation concepts

This is the first of two papers that provide an expository discussion of the basic structure of the asymptotic theory of M-estimators in dynamic nonlinear models and a revlew of the literature. In this paper we discuss consistency,uniform laws of large numbers and develop a new framework for laws of large numbers for dependent and heterogeneous processes, encompassing the theory of stochastically stable as well as near epoch dependent processes. This framework results in simplified catalogues of sufficient conditions for consistency. The second paper, Potscher and Prucha (1990b), deals with asymptotic distribution theory.     

Testing for unit roots in time series with nearly deterministic seasonal variation

This paper addresses the problem of testing for the presence of unit autoregressive roots in seasonal time series with negatively correlated moving average components. For such cases, many of the commonly used tests are known to have exact sizes much higher than their nominal significance level. We propose modifications of available test procedures that are based on suitably prewhitened data and feasible generalized least squares estimators. Monte Carlo experiments show that such modifications are successful in reducing size distortions in samples of moderate size.     

Simple least squares estimation versus best linear unbiased prediction

Necessary and sufficient conditions are developed for the simple least squares estimator to coincide with the best linear unbiased predictor. The conditions obtained are valid for a general linear model and are generalizations of the condition given by Watson (1972). Also, as a preliminary result, a new representation of the best linear unbiased predictor is established.     

Testing equality of ratios from two finite populations

Approximations are given for the bias and variance of both the regression and ratio estimator when sampling from a finite population, and simulation results are given indicating the accuracy of the approximations and the bias of the estimated approximations. A different estimator for the variance of the regression estimator is recommended. Test procedures are proposed for testing the hypothesis of equality of ratios from two finite populations, the procedures depending upon the types of populations being sampled. Simulation results indicating the effectiveness of the test procedures in controlling their size are given.     

Incidence-based estimates of life expectancy of the healthy for the UK: coherence between transition probabilities and aggregate life-tables

Summary.  Will the UK's aging population be fit and independent, or suffer from greater chronic ill health? Life expectancy of healthy people represents the expected number of years of healthy well-being that a life-table cohort would experience if age-specific rates of mortality and disability prevailed throughout the cohort's lifetime. Robust estimation of this life expectancy is thus essential for examining whether additional years of life are spent in good health and whether life expectancy is increasing faster than the decline of rates of disability. The paper examines a means of generating estimates of life expectancy for people who are healthy and unhealthy for the UK that are consistent with exogenous population mortality data. The method takes population transition matrices and adjusts these in a statistically coherent way so as to render them consistent with aggregate life-tables.     

Nonlinear models,residuals and transformations 1

A general review is given of statistical methods for nonlinear regression situations and some associated problems are mentioned. After general preliminary comments, a number of types of nonlinear model are outlined. Their parameterization is then examined. Analysis of residuals, with some emphasis on nonlinear models, is then considered and finally transformations of response and explanatory variable are discussed.     

Parametric Proportional Odds Frailty Models

We define a parametric proportional odds frailty model to describe lifetime data incorporating heterogeneity between individuals. An unobserved individual random effect, called frailty, acts multiplicatively on the odds of failure by time t. We investigate fitting by maximum likelihood and by least squares. For the latter, the parametric survivor function is fitted to the nonparametric Kaplan–Meier estimate at the observed failure times. Bootstrap standard errors and confidence intervals are obtained for the least squares estimates. The models are applied successfully to simulated data and to two real data sets. Least squares estimates appear to have smaller bias than maximum likelihood.