published_or_final_version / Mathematics / Master / Master of Philosophy

In longitudinal and spatio-temporal data analysis, repeated measurements from a subject can be either regional- or temporal-dependent. The correct specification of the within-subject covariance matrix cultivates an efficient estimation for mean regression coefficients. In this thesis, robust estimation for the mean and covariance jointly for the regression model of longitudinal data within the framework of generalized estimating equations (GEE) is developed. The proposed approach integrates the robust method and joint mean-covariance regression modeling. Robust generalized estimating equations using bounded scores and leverage-based weights are employed for the mean and covariance to achieve robustness against outliers. The resulting estimators are shown to be consistent and asymptotically normally distributed. Robust variable selection method in a joint mean and covariance model is considered, by proposing a set of penalized robust generalized estimating equations to estimate simultaneously the mean regression coefficients, the generalized autoregressive coefficients and innovation variances introduced by the modified Cholesky decomposition. The set of estimating equations select important covariate variables in both mean and covariance models together with the estimating procedure. Under some regularity conditions, the oracle property of the proposed robust variable selection method is developed. For these two robust joint mean and covariance models, simulation studies and a hormone data set analysis are carried out to assess and illustrate the small sample performance, which show that the proposed methods perform favorably by combining the robustifying and penalized estimating techniques together in the joint mean and covariance model. Capturing dynamic change of time-varying correlation structure is both interesting and scientifically important in spatio-temporal data analysis. The time-varying empirical estimator of the spatial correlation matrix is approximated by groups of selected basis matrices which represent substructures of the correlation matrix. After projecting the correlation structure matrix onto the space spanned by basis matrices, varying-coefficient model selection and estimation for signals associated with relevant basis matrices are incorporated. The unique feature of the proposed model and estimation is that time-dependent local region signals can be detected by the proposed penalized objective function. In theory, model selection consistency on detecting local signals is provided. The proposed method is illustrated through simulation studies and a functional magnetic resonance imaging (fMRI) data set from an attention deficit hyperactivity disorder (ADHD) study. / published_or_final_version / Statistics and Actuarial Science / Doctoral / Doctor of Philosophy

Thesis (Ph. D.)--Oregon State University, 2008. / Printout. Includes bibliographical references (leaves 64-66). Also available on the World Wide Web.

Background/Aims: Case-control designs are commonly employed in genetic association studies. In addition to the primary trait of interest, data on additional secondary traits, related to the primary trait, are often collected. Traditional association analyses between genetic variants and secondary traits can be biased in such cases, and several methods have been proposed to address this issue, including the inverse-probability-of-sampling-weighted (IPW) approach and semi-parametric maximum likelihood (SPML) approach. Methods: Here, we propose a set of new estimating equation based approach that combines observed and counter-factual outcomes to provide unbiased estimation of genetic associations with secondary traits. We extend the estimating equation framework to both generalized linear models (GLM) and non-parametric regressions, and compare it with the existing approaches. Results: We demonstrate analytically and numerically that our proposed approach provides robust and fairly efficient unbiased estimation in all simulations we consider. Unlike existing methods, it is less sensitive to the sampling scheme and underlying disease model specification. In addition, we illustrate our new approach using two real data examples. The first one is to analyze the binary secondary trait diabetes under GLM framework using a stroke case-control study. The second one is to analyze the continuous secondary trait serum IgE levels under linear and quantile regression models using an asthma case-control study. Conclusion: The proposed new estimating equation approach is able to accommodate a wide range of regressions, and it outperforms the existing approaches in some scenarios we consider.

This thesis aimed to apply the causal framework with potential outcomes to examine the causal effect of air temperature on reported violent crimes in Swedish municipalities. The Generalized Estimating Equations method was used on yearly, monthly and also July only data for the time period 2002-2014. One significant causal effect was established but the majority of the results pointed to there being no causal effect between air temperature and reported violent crimes.

Thesis (M.Phil.) - University of St Andrews, January 2008.

"This project utilized generalized estimating equations and general linear modeling to model learning curves for sonographer performance in emergency ultrasonography. Performance was measured in two ways: image quality (interpretable vs. possible hindrance in interpretation) and agreement of findings between the sonographer and an expert reviewing sonographer. Records from 109 sonographers were split into two data sets-- training (n=50) and testing (n=59)--to conduct exploratory analysis and fit the final models for analysis, respectively. We determined that the number of scans of a particular exam type required for a sonographer to obtain quality images on that exam type with a predicted probability of 0.9 is highly dependent upon the person conducting the review, the indication of the scan (educational or medical), and the outcome of the scan (whether there is a pathology positive finding). Constructing family-wise 95% confidence intervals for each exam type demonstrated a large amount of variation for the number of scans required both between exam types and within exam types. It was determined that a sonographer's experience with a particular exam type is not a significant predictor of future agreement on that exam type and thus no estimates were made based on the agreement learning curves. In addition, we concluded based on a type III analysis that when already considering exam type related experience, the consideration of experience on other exam types does not significantly impact the learning curve for quality. However, the learning curve for agreement is significantly impacted by the additional consideration of experience on other exam types."

Uma suposição muito comum na análise de modelos de regressão é a de respostas independentes. No entanto, quando trabalhamos com dados longitudinais ou agrupados essa suposição pode não fazer sentido. Para resolver esse problema existem diversas metodologias, e talvez a mais conhecida, no contexto não Gaussiano, é a metodologia de Equações de Estimação Generalizadas (EEGs), que possui similaridades com os Modelos Lineares Generalizados (MLGs). Essas similaridades envolvem a classificação do modelo em torno de distribuições da família exponencial e da especificação de uma função de variância. A única diferença é que nessa função também é inserida uma matriz trabalho que inclui a parametrização da estrutura de correlação dentro das unidades experimentais. O principal objetivo desta dissertação é estudar como esses modelos se comportam em uma situação específica, de dados de contagem com sobredispersão. Quando trabalhamos com MLGs esse problema é resolvido através do ajuste de um modelo com resposta binomial negativa (BN), e a ideia é a mesma para os modelos envolvendo EEGs. Essa dissertação visa rever as teorias existentes em EEGs no geral e para o caso específico quando a resposta marginal é BN, e além disso mostrar como essa metodologia se aplica na prática, com três exemplos diferentes de dados correlacionados com respostas de contagem. / An assumption that is common in the analysis of regression models is that of independent responses. However, when working with longitudinal or grouped data this assumption may not have sense. To solve this problem there are several methods, but perhaps the best known, in the non Gaussian context, is the one based on Generalized Estimating Equations (GEE), which has similarities with Generalized Linear Models (GLM). Such similarities involve the classification of the model around the exponential family and the specification of a variance function. The only diference is that in this function is also inserted a working correlation matrix concerning the correlations within the experimental units. The main objective of this dissertation is to study how these models behave in a specific situation, which is the one on count data with overdispersion. When we work with GLM this kind of problem is solved by setting a model with a negative binomial response (NB), and the idea is the same for the GEE methodology. This dissertation aims to review in general the GEE methodology and for the specific case when the responses follow marginal negative binomial distributions. In addition, we show how this methodology is applied in practice, with three examples of correlated data with count responses.

The sharing of needles among injection drug users (IDUs) is a common route of Human Immunodeficiency Virus and Hepatitis C Virus transmission. Through the increased utilization of social network analysis, researchers have been able to examine how the interpersonal relationships of IDUs affect injection risk behaviour. This study involves a secondary analysis of data from a cross-sectional study of 156 IDUs from Winnipeg, Manitoba titled “Social Network Analysis of Injection Drug Users”. Multiple logistic regression analysis was used to assess the individual and the social network characteristics associated with needle sharing among the IDUs. Generalized Estimating Equations analysis was used to determine the injecting dyad characteristics which influence needle sharing behaviour between the IDUs and their injection drug using network members. The results revealed five key thematic findings that were significantly associated with needle sharing: (1) types of drug use, (2) socio-demographic status, (3) injecting in semi-public locations, (4) intimacy, and (5) social influence. The findings from this study suggest that comprehensive prevention approaches that target individuals and their network relationships may be necessary for sustainable reductions in needle sharing among IDUs. / February 2006

Linear mixed models and generalized linear mixed models are random-effects models widely applied to analyze clustered or hierarchical data. Generally, random effects are often assumed to be normally distributed in the context of mixed models. However, in the mixed-effects logistic model, the violation of the assumption of normally distributed random effects may result in inconsistency for estimates of some fixed effects and the variance component of random effects when the variance of the random-effects distribution is large. On the other hand, summary statistics used for assessing goodness of fit in the ordinary logistic regression models may not be directly applicable to the mixed-effects logistic models. In this dissertation, we present our investigations of two independent studies related to goodness-of-fit tests in generalized linear mixed models. First, we consider a semi-nonparametric density representation for the random effects distribution and provide a formal statistical test for testing normality of the random-effects distribution in the mixed-effects logistic models. We obtain estimates of parameters by using a non-likelihood-based estimation procedure. Additionally, we not only evaluate the type I error rate of the proposed test statistic through asymptotic results, but also carry out a bootstrap hypothesis testing procedure to control the inflation of the type I error rate and to study the power performance of the proposed test statistic. Further, the methodology is illustrated by revisiting a case study in mental health. Second, to improve assessment of the model fit in the mixed-effects logistic models, we apply the nonparametric local polynomial smoothed residuals over within-cluster continuous covariates to the unweighted sum of squares statistic for assessing the goodness-of-fit of the logistic multilevel models. We perform a simulation study to evaluate the type I error rate and the power performance for detecting a missing quadratic or interaction term of fixed effects using the kernel smoothed unweighted sum of squares statistic based on the local polynomial smoothed residuals over x-space. We also use a real data set in clinical trials to illustrate this application.