Applied Statistics Commons^™

Unitary And Symmetric Structure In Deep Neural Networks, Kehelwala Dewage Gayan Maduranga

Theses and Dissertations--Mathematics

Recurrent neural networks (RNNs) have been successfully used on a wide range of sequential data problems. A well-known difficulty in using RNNs is the vanishing or exploding gradient problem. Recently, there have been several different RNN architectures that try to mitigate this issue by maintaining an orthogonal or unitary recurrent weight matrix. One such architecture is the scaled Cayley orthogonal recurrent neural network (scoRNN), which parameterizes the orthogonal recurrent weight matrix through a scaled Cayley transform. This parametrization contains a diagonal scaling matrix consisting of positive or negative one entries that can not be optimized by gradient descent. Thus the …

Orthogonal Recurrent Neural Networks And Batch Normalization In Deep Neural Networks, Kyle Eric Helfrich

Theses and Dissertations--Mathematics

Despite the recent success of various machine learning techniques, there are still numerous obstacles that must be overcome. One obstacle is known as the vanishing/exploding gradient problem. This problem refers to gradients that either become zero or unbounded. This is a well known problem that commonly occurs in Recurrent Neural Networks (RNNs). In this work we describe how this problem can be mitigated, establish three different architectures that are designed to avoid this issue, and derive update schemes for each architecture. Another portion of this work focuses on the often used technique of batch normalization. Although found to be successful …

Semiparametric And Nonparametric Methods For Comparing Biomarker Levels Between Groups, Yuntong Li

Theses and Dissertations--Statistics

Comparing the distribution of biomarker measurements between two groups under either an unpaired or paired design is a common goal in many biomarker studies. However, analyzing biomarker data is sometimes challenging because the data may not be normally distributed and contain a large fraction of zero values or missing values. Although several statistical methods have been proposed, they either require data normality assumption, or are inefficient. We proposed a novel two-part semiparametric method for data under an unpaired setting and a nonparametric method for data under a paired setting. The semiparametric method considers a two-part model, a logistic regression for …

Nonparametric Analysis Of Clustered And Multivariate Data, Yue Cui

Theses and Dissertations--Statistics

In this dissertation, we investigate three distinct but interrelated problems for nonparametric analysis of clustered data and multivariate data in pre-post factorial design.

In the first project, we propose a nonparametric approach for one-sample clustered data in pre-post intervention design. In particular, we consider the situation where for some clusters all members are only observed at either pre or post intervention but not both. This type of clustered data is referred to us as partially complete clustered data. Unlike most of its parametric counterparts, we do not assume specific models for data distributions, intra-cluster dependence structure or variability, in effect …

Nonparametric Tests Of Lack Of Fit For Multivariate Data, Yan Xu

Theses and Dissertations--Statistics

A common problem in regression analysis (linear or nonlinear) is assessing the lack-of-fit. Existing methods make parametric or semi-parametric assumptions to model the conditional mean or covariance matrices. In this dissertation, we propose fully nonparametric methods that make only additive error assumptions. Our nonparametric approach relies on ideas from nonparametric smoothing to reduce the test of association (lack-of-fit) problem into a nonparametric multivariate analysis of variance. A major problem that arises in this approach is that the key assumptions of independence and constant covariance matrix among the groups will be violated. As a result, the standard asymptotic theory is not …

Modelling Interactions Among Offenders: A Latent Space Approach For Interdependent Ego-Networks, Isabella Gollini, Alberto Caimo, Paolo Campana

Articles

Illegal markets are notoriously difficult to study. Police data offer an increasingly exploited source of evidence. However, their secondary nature poses challenges for researchers. A key issue is that researchers often have to deal with two sets of actors: targeted and non-targeted. This work develops a latent space model for interdependent ego-networks purposely created to deal with the targeted nature of police evidence. By treating targeted offenders as egos and their contacts as alters, the model (a) leverages on the full information available and (b) mirrors the specificity of the data collection strategy. The paper then applies this approach to …

Mass Grave Localization Prediction With Geographical Information Systems In Guatemala And Future Impacts, Perla Santillan

Master of Science in Forensic Science Directed Research Projects

Conducting physical searches for mass grave locations based on anecdotal evidence is a time consuming and resource intensive endeavor in circumstances that often pose a threat to personal safety. The development of tools and procedures to speed such searches can greatly reduce the risk involved, increase the number of individuals whose remains are recovered and identified; and, more importantly, reunite these remains with their loved ones to provide them with a proper burial. Geographic information systems (GIS) software, which can analyze and manipulate the spatial characteristics of known mass grave data, represents a powerful tool that can be used to …

Bayesian Kinetic Modeling For Tracer-Based Metabolomic Data, Xu Zhang

Theses and Dissertations--Statistics

Kinetic modeling of the time dependence of metabolite concentrations including the unstable isotope labeled species is an important approach to simulate metabolic pathway dynamics. It is also essential for quantitative metabolic flux analysis using tracer data. However, as the metabolic networks are complex including extensive compartmentation and interconnections, the parameter estimation for enzymes that catalyze individual reactions needed for kinetic modeling is challenging. As the pa- rameter space is large and multi-dimensional while kinetic data are comparatively sparse, the estimation procedure (especially the point estimation methods) often en- counters multiple local maximum such that standard maximum likelihood methods may yield …

Statistical Intervals For Various Distributions Based On Different Inference Methods, Yixuan Zou

Theses and Dissertations--Statistics

Statistical intervals (e.g., confidence, prediction, or tolerance) are widely used to quantify uncertainty, but complex settings can create challenges to obtain such intervals that possess the desired properties. My thesis will address diverse data settings and approaches that are shown empirically to have good performance. We first introduce a focused treatment on using a single-layer bootstrap calibration to improve the coverage probabilities of two-sided parametric tolerance intervals for non-normal distributions. We then turn to zero-inflated data, which are commonly found in, among other areas, pharmaceutical and quality control applications. However, the inference problem often becomes difficult in the presence of …

Measuring Change: Prediction Of Early Onset Sepsis, Aric Schadler

Theses and Dissertations--Statistics

Sepsis occurs in a patient when an infection enters into the blood stream and spreads throughout the body causing a cascading response from the immune system. Sepsis is one of the leading causes of morbidity and mortality in today’s hospitals. This is despite published and accepted guidelines for timely and appropriate interventions for septic patients. The largest barrier to applying these interventions is the early identification of septic patients. Early identification and treatment leads to better outcomes, shorter lengths of stay, and financial savings for healthcare institutions. In order to increase the lead time in recognizing patients trending towards septicemia …

Three Essays On Health Economics And Policy Evaluation, Shishir Shakya

Graduate Theses, Dissertations, and Problem Reports (ETD)

This dissertation consists of three essays on the U.S. Health care policy. Each paragraph below refers to the three abstracts for the three chapters in this dissertation, respectively. I provide quantitative evidence on how much Prescription Drug Monitoring Programs (PDMPs) affects the retail opioid prescribing behaviors. Using the American Community Survey (ACS), I retrieve county-level high dimensional panel data set from 2010 to 2017. I employ three separate identification strategies: difference-in-difference, double selection post-LASSO, and spatial difference-in-difference. I compare how the retail opioid prescribing behaviors of counties, that are mandatory for prescribers to check the PDMP before prescribing controlled substances …

Evaluating An Ordinal Output Using Data Modeling, Algorithmic Modeling, And Numerical Analysis, Martin Keagan Wynne Brown

Murray State Theses and Dissertations

Data and algorithmic modeling are two different approaches used in predictive analytics. The models discussed from these two approaches include the proportional odds logit model (POLR), the vector generalized linear model (VGLM), the classification and regression tree model (CART), and the random forests model (RF). Patterns in the data were analyzed using trigonometric polynomial approximations and Fast Fourier Transforms. Predictive modeling is used frequently in statistics and data science to find the relationship between the explanatory (input) variables and a response (output) variable. Both approaches prove advantageous in different cases depending on the data set. In our case, the data …

Elucidating The Properties And Mechanism For Cellulose Dissolution In Tetrabutylphosphonium-Based Ionic Liquids Using High Concentrations Of Water, Brad Crawford

Graduate Theses, Dissertations, and Problem Reports (ETD)

The structural, transport, and thermodynamic properties related to cellulose dissolution by tetrabutylphosphonium chloride (TBPCl) and tetrabutylphosphonium hydroxide (TBPH)-water mixtures have been calculated via molecular dynamics simulations. For both ionic liquid (IL)-water solutions, water veins begin to form between the TBPs interlocking arms at 80 mol % water, opening a pathway for the diffusion of the anions, cations, and water. The water veins allow for a diffusion regime shift in the concentration region from 80 to 92.5 mol % water, providing a higher probability of solvent interaction with the dissolving cellulose strand. The hydrogen bonding was compared between small and large …

Interpretations Of Bicoherence In Space & Lab Plasma Dynamics, Gregory Allen Riggs

Graduate Theses, Dissertations, and Problem Reports (ETD)

The application of bicoherence analysis to plasma research, particularly in non-linear, coupled-wave regimes, has thus far been significantly belied by poor resolution in time, and/or outright destruction of frequency information. Though the typical power spectrum cloaks the phase-coherency between frequencies, Fourier transforms of higher-order convolutions provide an n-dimensional spectrum which is adept at elucidating n-wave phase coherence. As such, this investigation focuses on the utility of the normalized bispectrum for detection of wave-wave coupling in general, with emphasis on distinct implications within the scope of non-linear plasma physics. Interpretations of bicoherent features are given for time series from …

The Analysis Of Neural Heterogeneity Through Mathematical And Statistical Methods, Kyle Wendling

Theses and Dissertations

Diversity of intrinsic neural attributes and network connections is known to exist in many areas of the brain and is thought to significantly affect neural coding. Recent theoretical and experimental work has argued that in uncoupled networks, coding is most accurate at intermediate levels of heterogeneity. I explore this phenomenon through two distinct approaches: a theoretical mathematical modeling approach and a data-driven statistical modeling approach.

Through the mathematical approach, I examine firing rate heterogeneity in a feedforward network of stochastic neural oscillators utilizing a high-dimensional model. The firing rate heterogeneity stems from two sources: intrinsic (different individual cells) and network …

Natural Lead-In Approaches To Response-Adaptive Allocation In Clinical Trials, Erin E. Donahue

Theses and Dissertations

Response-adaptive (RA) allocation designs can be implemented in clinical trials to skew the allocation of incoming subjects toward the better performing treatment group based on the previously accrued subjects' responses. These designs alleviate potential ethical concerns of equally allocating subjects in a trial when one treatment arm is inferior. The RA design can be generalized to include covariate information in the covariate-adjusted response-adaptive (CARA) design, which aims to maximize treatment successes conditional on a set of patient characteristics. While RA and CARA designs can improve the treatment of patients, they have unstable estimators and increased variability in early stages of …

Zero-Inflated Longitudinal Mixture Model For Stochastic Radiographic Lung Compositional Change Following Radiotherapy Of Lung Cancer, Viviana A. Rodríguez Romero

Theses and Dissertations

Compositional data (CD) is mostly analyzed as relative data, using ratios of components, and log-ratio transformations to be able to use known multivariable statistical methods. Therefore, CD where some components equal zero represent a problem. Furthermore, when the data is measured longitudinally, observations are spatially related and appear to come from a mixture population, the analysis becomes highly complex. For this matter, a two-part model was proposed to deal with structural zeros in longitudinal CD using a mixed-effects model. Furthermore, the model has been extended to the case where the non-zero components of the vector might a two component mixture …

Tropical Cyclone Hazards In Relation To Propagation Speed, Jiehao Huang

Dissertations and Theses

As the population and infrastructure along the US East Coast increase, it becomes increasingly important to study the characteristics of tropical cyclones that can impact the coast. A recent study shows that the propagation speed of tropical cyclones has slowed over the past 60 years, which can lead to greater accumulation of precipitation and greater storm surge impacts. The study presented herein is meant to examine and analyze the relationships that exist between the propagation speed of tropical cyclones, their surface wind strength, displacement angles, and cyclone averaged winds. This analysis is focused on tropical cyclones spanning from 1950-2015 in …

Statistical Analysis Of Demographic Effects On Insurance Coverage Of Perinatal And Neonatal Morbidity, Madeline Durbin

Undergraduate Honors Thesis Projects

In the United States of America, Ohio has one of the worst neonatal and perinatal death rates. Within Ohio, Montgomery County has an above average neonatal and perinatal death rate. This statistic can be lowered if more women in Montgomery County have health insurance. They would be more likely to seek out prenatal health care, since they would no longer have to pay as much money out-of-pocket. This would allow medical professionals to be able to diagnose and treat any potential issues in the mother or child earlier. Having health insurance would also prevent mothers-to-be from seeking out other potentially …

Phenotype Extraction: Estimation And Biometrical Genetic Analysis Of Individual Dynamics, Kevin L. Mckee

Theses and Dissertations

Within-person data can exhibit a virtually limitless variety of statistical patterns, but it can be difficult to distinguish meaningful features from statistical artifacts. Studies of complex traits have previously used genetic signals like twin-based heritability to distinguish between the two. This dissertation is a collection of studies applying state-space modeling to conceptualize and estimate novel phenotypic constructs for use in psychiatric research and further biometrical genetic analysis. The aims are to: (1) relate control theoretic concepts to health-related phenotypes; (2) design statistical models that formally define those phenotypes; (3) estimate individual phenotypic values from time series data; (4) consider hierarchical …