Thesis (Ph. D.)--University of Washington, 2003. / Vita. Includes bibliographical references (leaves 64-67).

No description available.

Not only basic research in sciences, but also medicine, law, and manufacturing need statistical techniques, including graphics, to assess disagreement. For some items or individuals ⍳ = 1,2,---,ո suppose that pairs (X⍳,Y⍳) denote each item's measurements by two distinct methods or by two observers, or X⍳ and Y⍳ may be initial and repeat measurement scores, with discrepancy D⍳ = X⍳ - Y⍳. Disagreement may be characterized by location and scale parameters of discrepancy distributions. The present work primarily addresses estimation of central tendency - relative bias or median discrepancy (or discrepancy rate in some instances). Most previous literature on "agreement" or "reliability" instead concerns X, Y correlation, which can be regarded as the complement of discrepancy variance. (There is ambiguity or confusion about concepts of "reliability" in the literature of various applications.) Discrepancies D₁, D₂, • • •, Dո in practice often violate assumptions of standard statistical models and methods that have been commonly applied in studies of agreement. In particular, both X⍳ and Y⍳ generally incorporate measurement errors. Further, these two measurement error distributions for the ⍳th item need not be the same; and both distributions could depend on the magnitude µ⍳, of the item being measured. Hence, for example, discrepancy D⍳ could have variance proportional to the size of the item; and in general D₁, D₂, • • •, Dո are not identically distributed. Finally, the selection of items ⍳ = 1,2, • • •, ո often is not random. To estimate median discrepancy, we consider nonparametric confidence intervals corresponding to Student t test, sign test, Wilcoxon signed rank test, or other permutation tests. Several criteria are developed to compare the performance of one procedure relative to another, including expected ratio of confidence interval lengths (related to Pitman asymptotic relative efficiency of tests) and relative variability of interval lengths. Theoretical calculations and Monte Carlo simulation results suggest different procedural preferences for random sampling from different distributions. For discrepancies distributed non-identically, but symmetrically about a common median value, mixture sampling is used as an approximate model. This approach is related to a "random walk" (rather than random sample) model of D₁, D₂, • • •, Dո proposed particularly for discrepancies between counting processes. We also emphasize graphic methods, especially plots of difference of Y - X versus average (X + Y)/2, for exploratory analysis of discrepancy data and to choose appropriate statistical models and numerical methods. Various data sets are analyzed as examples of the methodology. / Science, Faculty of / Statistics, Department of / Graduate

Thesis (Ph. D.)--University of Chicago, Dept. of Economics, June 2000. / Includes bibliographical references. Also available on the Internet.

Thesis (M.Sc.). / Written for the School of Computer Science. Title from title page of PDF (viewed 2009/06/30). Includes bibliographical references.

Thesis (Ph. D.)--University of California, San Diego, 2006. / Title from first page of PDF file (viewed September 20, 2006). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references.

M.Sc. / In this study we consider the problem ofestiniating the slope in the simple linear errors-in-variables model. There are two different types of relationship that can he specified in the errors-in-variables model: one that specifies a functional linear relationship and one describing a structural linear relationship. The different relationship specifications can lead to different estimators with different properties. These two specifications are highlighted in this study. A least squares solution (to the estimation of the slope) is given. The problem of finding the maximum likelihood solution to these two specifications is addressed. It is noted that an unidentifiability problem arises in this attempt. The solution is seen to lie in making assumptions on the error variances. Interval estimation for the slope parameter is discussed. It is noted that any interval estimator of the slope whose length is always finite will have a confidence coefficient of zero. Various interval estimation methods are reviewed but emphasis is mainly on the investigation of a bootstrap procedure for estimating the confidence interval for the slope parameter β. More specifically, the Linder and Babu (1994) (bootstrap) method for the structural relationship model with known variance ratio is investigated here. The error distributions were assumed normal. A simulation study based on this paper is carried out. The results in the simulation study show that this bootstrap procedure performs well in comparison with the normal theory estimates for normally distributed data, that is, it has better coverage accuracy than the normal approximation.

High-dimensional data, such as that relating to gene expression in microarray experiments, may contain substantial amount of useful information to be explored. However, the information, relevant variables and their joint interactions are usually diluted by noise due to a large number of non-informative variables. Consequently, variable selection plays a pivotal role for learning in high dimensional problems. Most of the traditional feature selection methods, such as Pearson's correlation between response and predictors, stepwise linear regressions and LASSO are among the popular linear methods. These methods are effective in identifying linear marginal effect but are limited in detecting non-linear or higher order interaction effects. It is well known that epistasis (gene - gene interactions) may play an important role in gene expression where unknown functional forms are difficult to identify. In this thesis, we propose a novel nonparametric measure to first screen and do feature selection based on information from nearest neighborhoods. The method is inspired by Lo and Zheng's earlier work (2002) on detecting interactions for discrete predictors. We apply a backward elimination algorithm based on this measure which leads to the identification of many in influential clusters of variables. Those identified groups of variables can capture both marginal and interactive effects. Second, each identified cluster has the potential to perform predictions and classifications more accurately. We also study procedures how to combine these groups of individual classifiers to form a final predictor. Through simulation and real data analysis, the proposed measure is capable of identifying important variable sets and patterns including higher-order interaction sets. The proposed procedure outperforms existing methods in three different microarray datasets. Moreover, the nonparametric measure is quite flexible and can be easily extended and applied to other areas of high-dimensional data and studies.

Ma Ning. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (leaves 49-57). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.5 / Chapter 2 --- Literature Review --- p.11 / Chapter 2.1 --- Problems about Using Sibling Samples --- p.19 / Chapter 2.2 --- Difficulties with the Within-twin-pair Studies --- p.20 / Chapter 3 --- Method --- p.21 / Chapter 3.1 --- Omitted Variable Bias (Selection Effect) --- p.21 / Chapter 3.1.1 --- OLS Model --- p.21 / Chapter 3.1.2 --- Fixed-Effect Model --- p.23 / Chapter 3.1.3 --- GLS Model --- p.23 / Chapter 3.2 --- Measurement Error --- p.24 / Chapter 4 --- Data --- p.26 / Chapter 5 --- Results --- p.29 / Chapter 5.1 --- "OLS, Fixed-Effect, GLS and IV estimates" --- p.29 / Chapter 5.2 --- Important findings --- p.34 / Chapter 5.3 --- Further Results --- p.35 / Chapter 5.3.1 --- Consistency of Fixed-Effect Estimate --- p.35 / Chapter 5.3.2 --- Smoking as an Instrument for Education --- p.39 / Chapter 5.3.3 --- Symmetry Test --- p.41 / Chapter 5.3.4 --- Hausman Test --- p.44 / Chapter 5.3.5 --- Selection Bias --- p.45 / Chapter 6 --- Conclusions --- p.48 / Chapter 7 --- Bibliography --- p.49

Orientador: Luiz Koodi Hotta / Dissertação ( mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Computação Cientifica / Made available in DSpace on 2018-08-15T13:05:45Z (GMT). No. of bitstreams: 1 Ribeiro_AndreLuizPrima_M.pdf: 3151695 bytes, checksum: d87230fa6191977394ccb585657639ad (MD5) Previous issue date: 2010 / Resumo: O conhecimento das relações de dependência entre as economias são relevantes para tomadas de decisões de Bancos Centrais, investidores e governos. Um tema desafiador é o estudo da existência de contágio entre as economias. Este trabalho considera o Modelo Canônico de Contágio estudado por Pesaran e Pick (2007), o qual diferencia contágio de interdependência. O estimador de mínimos quadrados ordinário para este modelo é viesado devido à existência de variáveis endógenas no modelo. A teoria de variáveis instrumentais é utilizada para diminuir o viés existente nos estimadores de mínimos quadrados ordinários. Este trabalho estuda este modelo na presença de erros heteroscedásticos e utiliza as volatilidades condicionais como variáveis instrumentais. São estudados vários métodos para teste de hipóteses, com ênfase em testes robustos a instrumentos fracos. São abordadas duas diferentes definições de crise e são postuladas como instrumentos válidos as volatilidades condicionais dos índices de desempenho das economias e analisadas por meio de simulações de Monte Carlo a validade destes instrumentos para identificar a existência de contágio. Especificamente, são consideradas as distribuições dos estimadores e a função poder dos testes propostos para diferentes tamanhos de amostras, bem como, estudadas as aproximações das distribuições assintóticas dos estimadores e estatísticas dos testes. Finalmente, o modelo canônico de contágio é utilizado na análise dos dados de retorno dos principais índices acionários de Argentina, Brasil, México e EUA, assim como para alguns países asiáticos / Abstract: The understanding of the dependence among the economies are relevant to policy makers, Central Banks and investors in the decision making process. An important issue is the study of the existence of contagion among the economies. This work consider the Canonical Model of Contagion of Pesaran and Pick (2007), which diferentiates contagion of interdependence. The ordinary least squares estimator for this model is biased because there are endogenous variables in the model. Instrumental variable are used in order to decrease the bias of the ordinary least squares estimators. The model is extended to the case of heteroskedastic errors, feature usually found in financial data. Two definitions of crises are applied and we postulate the conditional volatility of the performance indexes as a instrumental variable. We analyze the validity of this instruments by means of Monte Carlo simulations. Monte Carlo simulations are used to analyst the distributions of the estimators and the power functions of the tests proposed. Finally, the canonical model of contagion is used to analyst the data of the most important performance indexes of Argentina, Brazil, Mexico and USA, as well the performance indexes of seven Asiatic countries / Mestrado / Estatistica / Mestre em Estatística