Statistics and Probability Commons^™

Semi-Parametrics Dose Finding Methods, Matthieu Clertant, John O'Quigley

COBRA Preprint Series

We describe a new class of dose finding methods to be used in early phase clinical trials. Under some added parametric conditions the class reduces to the family of continual reassessment method (CRM) designs. Under some relaxation of the underlying structure the method is equivalent to the CCD, mTPI or BOIN classes of designs. These latter designs are non-parametric in nature whereas the CRM class can be viewed as being strongly parametric. The proposed class is characterized as being semi-parametric since it corresponds to CRM with a nuisance parameter. Performance is good, matching that of the CRM class and improving …

The Association Between Osteoporosis And Early Menopause Following Hysterectomy, Mia Meeyaong-Won Botkin

Walden Dissertations and Doctoral Studies

Osteoporosis is considered to be the most adverse public health disease associated with substantial mortality among postmenopausal women. Hysterectomy, surgically induced menopause, contributes to the early onset of menopause. However, there was no evidence of an association between early menopause following hysterectomy and osteoporosis among postmenopausal women. The purpose of this quantitative study was to examine the association between demographic and behavioral factors and the prevalence of osteoporosis among hysterectomized postmenopausal women. The integrated theory of health behavior change theoretical framework guided study. Cross-sectional secondary data from the 2009-2010 National Health and Nutrition Examination Survey were used. Multiple logistic regression …

Space-Time Modelling Of Emerging Infectious Diseases: Assessing Leptospirosis Risk In Sri Lanka, Cameron C F Plouffe

Theses and Dissertations (Comprehensive)

In this research, models were developed to analyze leptospirosis incidence in Sri Lanka and its relation to rainfall. Before any leptospirosis risk models were developed, rainfall data were evaluated from an agro-ecological monitoring network for producing maps of total monthly rainfall in Sri Lanka. Four spatial interpolation techniques were compared: inverse distance weighting, thin-plate splines, ordinary kriging, and Bayesian kriging. Error metrics were used to validate interpolations against independent data. Satellite data were used to assess the spatial pattern of rainfall. Results indicated that Bayesian kriging and splines performed best in low and high rainfall, respectively. Rainfall maps generated from …

Score Test Derivations And Implementations For Bivariate Probability Mass And Density Functions With An Application To Copula Functions, Roy Bower

Theses and Dissertations

This dissertation is comprised and grounded in statistical theory with an application to solving real world problems. In particular, the development and implementation of multiple score tests under a variety of scenarios are derived, applied, and interpreted. In chapter 2, I propose a score test for independence of the marginals based on Lakshminarayana’s bivariate Poisson distribution. Each marginal distribution of the bivariate model is a univariate Poisson distribution, and the parameters of the bivariate distribution can be estimated using maximum likelihood methods. The simulation study shows that the score test maintains size close to the nominal level. To assess the …

Bayesian Ensemble Of Regression Trees For Multinomial Probit And Quantile Regression, Bereket P. Kindo

Theses and Dissertations

This dissertation proposes multinomial probit Bayesian additive regression trees (MPBART), ordered multiclass Bayesian additive classification trees (O-MBACT) and Bayesian quantile additive regression trees (BayesQArt) as extensions of BART - Bayesian additive regression trees for tackling multinomial choice, multiclass classification, ordinal regression and quantile regression problems. The proposed models exhibit very good predictive performances. In particular, ranking among the top performing procedures when non-linear relationships exist between the response and the predictors. The proposed procedures can readily be applied on data sets with the number of predictors larger than the number of observations.

MPBART is sufficiently flexible to allow inclusion of …

Semiparametric Estimation Methods For Complex Accelerated Failure Time Model, Yinding Wang

Theses and Dissertations

The proportional hazards (PH) model and the accelerated failure time (AFT) model are the two most popular survival models in fitting the right-censored data. The AFT model is a useful alternative to the PH model, particularly when the PH assumption is not satisfied. Usually, the linear association is assumed with logarithm of survival time in the AFT model. However, the nonlinear association may exist in practice. The first project aims to handle the nonlinear component in the AFT model, which is called the semiparametric additive partial accelerated failure time (AP-AFT) model. Two estimation methods based on the rank-smooth method and …

Sample Size Calculation For Ph Mixture Cure Model, Yihong Zhan

Theses and Dissertations

With the development of advanced medical technology, a significant proportion of patients can be cured of many chronic diseases. Because a substantial fraction of patients have censored information, the standard survival model, such as the proportional hazards (PH) model cannot capture the cured information of patients. Thus PH mixture cure model is developed to handle the survival data with potential cured information. A corresponding sample size formula based on log rank test has been proposed by Wang et al. (2012) and the probability of death in their formula is only contributed by the control arm. However, to calculate the sample …

Registration And Clustering Of Functional Observations, Zizhen Wu

Theses and Dissertations

As an important exploratory analysis, curves of similar shape are often classified into groups, which we call clustering of functional data. Phase variations or time distortions are often encountered in the biological processes, such as growth patterns or gene profiles. As a result of time distortion, curves of similar shape may not be aligned. Regular clustering methods for functional data usually ignore the presence of phase variations, which may result in low clustering accuracy. However, it is difficult to account for phase variation without knowing the cluster structure.

In this dissertation, we first propose a Bayesian method that simultaneously clusters …

Modern Estimation Problems In Group Testing, Md Shamim Sarker

Theses and Dissertations

In the simplest form of group testing, pools are formed by compositing a fixed number of individual specimens (e.g., blood, urine, swab, etc.) and then the pools are tested for a binary characteristic, such as presence or absence of a disease. Group testing is commonly used to screen for a variety of sexually transmitted diseases in epidemiological applications where the main goal is to increase testing efficiency. In this dissertation, we study three estimation problems that are motivated by real-life applications. We propose new methods to model group testing data for both single and multiple infections. In the first problem, …

Spatio-Temporal Analysis Of The Occupational Fatal Victimization Of Law Enforcement Officers In The Us, Xueyi Xing

Theses and Dissertations

The models with constant coefficients of the covariates across space and time are commonly used in spatio-temporal analyses. However, the associations between risk factors and the outcome could have locally differential temporal trends in many cases. In this study, a Bayesian latent cluster modeling strategy is employed to identify potential spatial clusters in which locally specific sets of temporally varying coefficients of covariates are allowed. A state-level panel data of police officers occupational fatal victimization for the years 1979-2010 is used. To accommodate overdisperson and excess zeros, a negative binomial model and zero-inflated Poisson/negative binomial models are also utilized. A …

Regression Models For Count Data Based On The Double Poisson Distribution, Rebecca Wardrop

Theses and Dissertations

This paper explores the double Poisson distribution. The probability mass function and the difficulties associated with derivative-based optimization for this distribution are discussed. Stata software developed for estimation of double Poisson regression is detailed. Simulations are used to test the software. Data which are over-, under-, and equidispersed relative to the Poisson are generated and the software is utilized to estimate a regression model, a zero-inflated model, and a marginalized zero-inflated model all based on the double Poisson distribution. The estimated power of the test for φ = 1 for the double Poisson models are compared to the power of …

An Online Statistics Course From Faculty And Students' Perspectives: A Case Study, Ruth Best

Walden Dissertations and Doctoral Studies

Faculty at a private college in the northeastern United States found students lacked prerequisite mathematical skills and were unable to transfer quantitative reasoning skills to upper level business courses. Guided by Mezirow's transformative learning theory and Knowles' approach to self-directed learning, this study examined how undergraduate students learn statistics online. The purpose of this qualitative embedded case study was to examine faculty and students' perspectives about the online statistics course design and delivery while exploring possible barriers to students' learning. Data collection occurred by review of course documents and the learning management system. Archival data generated questions for semistructured interviews …

A Data Science Course For Undergraduates: Thinking With Data, Benjamin Baumer

Mathematics Sciences: Faculty Publications

Data science is an emerging interdisciplinary field that combines elements of mathematics, statistics, computer science, and knowledge in a particular application domain for the purpose of extracting meaningful information from the increasingly sophisticated array of data available in many settings. These data tend to be nontraditional, in the sense that they are often live, large, complex, and/or messy. A first course in statistics at the undergraduate level typically introduces students to a variety of techniques to analyze small, neat, and clean datasets. However, whether they pursue more formal training in statistics or not, many of these students will end up …

System-Wide Prediction Of General, All-Cause, Preventable Hospital Readmissions, Ken Musselman, Brandon Pope, Steve Witz, Zhiyi Tian, Lingsong Zhang, Linda Leon, Ann Davis

RCHE Publications

Existing studies of hospital readmissions typically focus on specific diagnoses, age groups, discharge dispositions, payer classes, or hospitals, and often use small samples. It is not clear how predictive models generated from such studies generalize across diseases, hospitals, or time periods. In this study, a logistic regression model of readmission risk within 30 days based on hospital administrative data was constructed and validated across hospitals and time periods. The hospitals included both general and specialty hospitals such as long-term care, women’s, and children’s hospitals. The administrative data included information on patient’s demographics, diagnoses, procedures, and discharge disposition. Derivation and validation …

Statistical Handling Of Medical Data - An Ethical Perspective, Ajay Kumar Bansal Dr

COBRA Preprint Series

Medical Science is a delicate subject and the clinical data generated from the medical trials must be reliable and of good quality. Not only the quality of generated data is important, but the management is also crucial and is to be handled very carefully. In this paper, the ethical aspect of statistical handling of such data is discussed.

Every profession has some set of norms to follow to achieve its objectives. These norms are called professional ethics which shows the essence of human behaviour. Same way, the field of medical research is expected to follow ethical norms, to obtain reliable …

Semi-Parametric Estimation And Inference For The Mean Outcome Of The Single Time-Point Intervention In A Causally Connected Population, Oleg Sofrygin, Mark J. Van Der Laan

U.C. Berkeley Division of Biostatistics Working Paper Series

We study the framework for semi-parametric estimation and statistical inference for the sample average treatment-specific mean effects in observational settings where data are collected on a single network of connected units (e.g., in the presence of interference or spillover). Despite recent advances, many of the current statistical methods rely on estimation techniques that assume a particular parametric model for the outcome, even though some of the most important statistical assumptions required by these models are most likely violated in the observational network settings, often resulting in invalid and anti-conservative statistical inference. In this manuscript, we rely on the recent methodological …

Statistical Estimation Of White Matter Microstructure From Conventional Mri, Leah Suttner, Amanda Mejia, Blake Dewey, Pascal Sati, Daniel S. Reich, Russell T. Shinohara

UPenn Biostatistics Working Papers

Diffusion tensor imaging (DTI) has become the predominant modality for studying white matter integrity in multiple sclerosis (MS) and other neurological disorders. Unfortunately, the use of DTI-based biomarkers in large multi-center studies is hindered by systematic biases that confound the study of disease-related changes. Furthermore, the site-to-site variability in multi-center studies is significantly higher for DTI than that for conventional MRI-based markers. In our study, we apply the Quantitative MR Estimation Employing Normalization (QuEEN) model to estimate the four DTI measures: MD, FA, RD, and AD. QuEEN uses a voxel-wise generalized additive regression model to relate the normalized intensities of …

Correction Of Verication Bias Using Log-Linear Models For A Single Binaryscale Diagnostic Tests, Haresh Rochani, Hani M. Samawi, Robert L. Vogel, Jingjing Yin

Biostatistics: Faculty Publications

In diagnostic medicine, the test that determines the true disease status without an error is referred to as the gold standard. Even when a gold standard exists, it is extremely difficult to verify each patient due to the issues of costeffectiveness and invasive nature of the procedures. In practice some of the patients with test results are not selected for verification of the disease status which results in verification bias for diagnostic tests. The ability of the diagnostic test to correctly identify the patients with and without the disease can be evaluated by measures such as sensitivity, specificity and predictive …

Applying Bayesian Machine Learning Methods To Theoretical Surface Science, Shane Carr

McKelvey School of Engineering Graduate Student Theses & Dissertations

Machine learning is a rapidly evolving field in computer science with increasingly many applications to other domains. In this thesis, I present a Bayesian machine learning approach to solving a problem in theoretical surface science: calculating the preferred active site on a catalyst surface for a given adsorbate molecule. I formulate the problem as a low-dimensional objective function. I show how the objective function can be approximated into a certain confidence interval using just one iteration of the self-consistent field (SCF) loop in density functional theory (DFT). I then use Bayesian optimization to perform a global search for the solution. …

A Generally Efficient Targeted Minimum Loss Based Estimator, Mark J. Van Der Laan

U.C. Berkeley Division of Biostatistics Working Paper Series

Suppose we observe n independent and identically distributed observations of a finite dimensional bounded random variable. This article is concerned with the construction of an efficient targeted minimum loss-based estimator (TMLE) of a pathwise differentiable target parameter based on a realistic statistical model.

The canonical gradient of the target parameter at a particular data distribution will depend on the data distribution through an infinite dimensional nuisance parameter which can be defined as the minimizer of the expectation of a loss function (e.g., log-likelihood loss). For many models and target parameters the nuisance parameter can be split up in two components, …