Statistics and Probability Commons^™

Sparse Partitioned Empirical Bayes Ecm Algorithms For High-Dimensional Linear Mixed Effects And Heteroscedastic Regression, Anja Zgodic

Theses and Dissertations

Variable selection methods in both the frequentist and Bayesian frameworks are powerful techniques that provide prediction and inference in high-dimensional linear regression models. These methods often assume independence between observations and normally distributed errors with the same variance. In practice, these two assumptions are often violated. To mitigate this, we develop efficient and powerful Bayesian approaches for linear mixed modeling and heteroscedastic linear regression. These method offers increased flexibility through the development of empirical Bayes estimators for hyperparameters, with computationally efficient estimation through the Expectation Conditional-Minimization (ECM) algorithm. The novelty of these approaches lies in the partitioning and parameter expansion, …

Wernicke's Encephalopathy: Mapping The Risk Factors Throughout The State Of South Carolina, Shannon M. Rychener

Senior Theses

Wernicke’s Encephalopathy is a consistently underrecognized neurodegenerative brain disorder resulting from prolonged thiamine deficiency. Clinical presentation of the disease results from brain lesions attributable to thiamine deficiency. Because these lesions occur in various locations in the cerebral cortex, symptoms can vary significantly. Varied presentation of symptoms, in addition to the lack of a widely accepted biomarker for the disorder cause challenges to clinicians when identifying and diagnosing the disorder. Due to these challenges, healthcare providers must heavily rely on patient history and risk factor prevalence when multiple symptoms of the disorder are present. By mapping the prevalence of the four …

Statistical Clustering Of Networks With Additional Information, Paul Atandoh

Dissertations

As the online market grows rapidly, many companies and researchers are interested in analyzing product review dataset which includes ratings and text review data. In the first project, we mainly focus on analyzing the text review data. In the current literature, it is common to use only text analysis tools to analyze review dataset. But in our work, we propose a method that utilizes both a text analysis method such as topic modeling and a statistical network model to build network among individuals and find interesting communities. We introduce a promising framework that incorporates topic modeling technique to define the …

High-Dimensional Variable Selection Via Knockoffs Using Gradient Boosting, Amr Essam Mohamed

Dissertations

As data continue to grow rapidly in size and complexity, efficient and effective statistical methods are needed to detect the important variables/features. Variable selection is one of the most crucial problems in statistical applications. This problem arises when one wants to model the relationship between the response and the predictors. The goal is to reduce the number of variables to a minimal set of explanatory variables that are truly associated with the response of interest to improve the model accuracy. Effectively choosing the true influential variables and controlling the False Discovery Rate (FDR) without sacrificing power has been a challenge …

That’S My Deity: An Examination Of Online Lokean Cultures Through Log-Linear Modeling, Mary Bernstein

Senior Theses

A rise in online religious communities and the growth of so-called ‘Old World’ religions are reflected in the internet’s subcultures of Neopaganism, a growing religious movement that has been documented in America since the 1960s. The religions under this umbrella movement vary drastically and include belief systems such as Wicca, Druidry, and deity worship. Belief systems under this movement lack the traditional hierarchy found in structured religion and lack a singular sacred text. As such, believers usually find and support one another not through a physical sacred place of meeting, but through an online community that acts as sacred space. …

Moral Injury To Inform Analysis Of Post-Traumatic Stress Disorder, Amanda Julia Manea

Senior Theses

Post-traumatic stress disorder (PTSD) is a mental health condition that almost one out of ten veterans struggle with. Although the National Center for PTSD has made extensive progress in characterizing and developing new treatments for PTSD, most veterans still experience symptoms of PTSD following treatment. Novel avenues of investigation, such as developing algorithms to review electronic health record (EHR) data and better understanding moral injury, are being pursued to address the gap that still exists when it comes to treating veterans. Moral injury is the individual evaluation of exposure to a potentially morally injurious event (PMIE) and can lead to …

Arousing Motives Or Eliciting Stories? On The Role Of Pictures In A Picture–Story Exercise, Philipp Schäpers, Stefan Krumm, Filip Lievens, Nikola Stenzel

Research Collection Lee Kong Chian School Of Business

Picture–story exercises (PSE) form a popular measurement approach that has been widely used for the assessment of implicit motives. However, current theorizing offers two diverging perspectives on the role of pictures in PSEs: either to elicit stories or to arouse motives. In the current study, we tested these perspectives in an experimental design. We administered a PSE either with or without pictures. Results from N = 281 participants revealed that the experimental manipulation had a medium to large effect for the affiliation and power motive domains, but no effect for the achievement motive domain. We conclude that the herein chosen …

Dynamic Equations, Control Problems On Time Scales, And Chaotic Systems, Martin Bohner

Mathematics and Statistics Faculty Research & Creative Works

The unification of integral and differential calculus with the calculus of finite differences has been rendered possible by providing a formal structure to study hybrid discrete-continuous dynamical systems besides offering applications in diverse fields that require simultaneous modeling of discrete and continuous data concerning dynamic equations on time scales. Therefore, the theory of time scales provides a unification between the calculus of the theory of difference equations with the theory of differential equations. In addition, it has become possible to examine diverse application problems more precisely by the use of dynamical systems on time scales whose calculus is made up …

Exploring Time-Varying Extraneous Variables Effects In Single-Case Studies, Ke Cheng

USF Tampa Graduate Theses and Dissertations

The effect of time-varying extraneous variables has been studied in other statistical analyses such as using Kaplan–Meier or Cox regression analysis in survival analyses. Nonetheless, the effect of modeling versus not modeling individual specific time varying extraneous variables has not been explored in multiple-baseline single case designs through Monte Carlo simulation studies. Therefore, in my dissertation, I used simulation methods to explore for a variety of conditions (varying in the number of participants, number of observations per participant, type of extraneous variable effect, size of the true intervention effect) the impact of extraneous variables on bias and standard error of …

Beyond Machine Learning: An Fmri Domain Adaptation Model For Multi-Study Integration, Lauryn Michelle Burleigh

LSU Doctoral Dissertations

Traditional machine learning analyses are challenging with functional magnetic
resonance imaging (fMRI) data, not only because of the amount of data that needs to be
collected, adding a particular challenge for human fMRI research, but also due to the change in
hypothesis being addressed with various analytical techniques. Domain adaptation is a type of
transfer learning, a step beyond machine learning which allows for multiple related, but not
identical, data to contribute to a model, can be beneficial to overcome the limitation of data
needed but may address different hypothesis questions than anticipated given the analysis
computation. This dissertation assesses …

Integrating And Optimizing Genomic, Weather, And Secondary Trait Data For Multiclass Classification, Vamsi Manthena, Diego Jarquín, Reka Howard

Department of Statistics: Faculty Publications

Modern plant breeding programs collect several data types such as weather, images, and secondary or associated traits besides the main trait (e.g., grain yield). Genomic data is high-dimensional and often over-crowds smaller data types when naively combined to explain the response variable. There is a need to develop methods able to effectively combine different data types of differing sizes to improve predictions. Additionally, in the face of changing climate conditions, there is a need to develop methods able to effectively combine weather information with genotype data to predict the performance of lines better. In this work, we develop a novel …

Using Physics-Informed Neural Networks For Multigrid In Time Coarse Grid Equations, Jonathan P. Gutierrez

Mathematics & Statistics ETDs

For parallel-in-time integration methods, the multigrid-reduction-in-time (MGRIT) method has shown promising results in both improved convergence and increased computational speeds when solving evolution problems. However, one problem the MGRIT algorithm currently faces is it struggles solving hyperbolic problems efficiently. In particular, hyperbolic problems are generally solved using explicit methods and this causes issues on the coarser multigrid levels, where larger (coarser) time step sizes can violate the stability condition. In this thesis, physics-informed neural networks (PINNs) are used to evaluate the coarse grid equations in the MGRIT algorithm with the goal to improve convergence for problems with hyperbolic behavior, as …

Statistical Analysis Of Ribonucleotide Incorporation In Human Cells, Tejasvi Channagiri

USF Tampa Graduate Theses and Dissertations

During the DNA replication process, ribonucleotides, the building blocks of RNA, may be occasionally incorporated in the newly synthesized DNA. DNA is primarily composed of deoxyribonucleotides and there exist cellular mechanisms for removing ribonucleotides from DNA, which may point towards ribonucleotide incorporation being a replication error. Further, an excess of these ribonucleotides in the genome has been known to lead to genomic instability and has been implicated in human diseases. However, there are also hypotheses that suggest that ribonucleotides may be beneficial in certain circumstances. In this study we examine ribonucleotide incorporation in the human genome in several human cell …

The Effects Of Demographics And Risk Factors On The Morphological Characteristics Of Human Femoropopliteal Arteries, Sayed Ahmadreza Razian, Majid Jadidi, Alexey Kamenskiy

UNO Student Research and Creative Activity Fair

Background: Disease of the lower extremity arteries (Peripheral Arterial Disease, PAD) is associated with high morbidity and mortality. During disease development, the arteries adapt by changing their diameter, wall thickness, and residual deformations, but the effects of demographics and risk factors on this process are not clear.

Methods: Superficial femoral arteries from 736 subjects (505 male, 231 female, 12 to 99 years old, average age 51±17.8 years) and the associated demographic and risk factor variables were used to construct machine learning (ML) regression models that predicted morphological characteristics (diameter, wall thickness, and longitudinal opening angle resulting from the …

Fraud Pattern Detection For Nft Markets, Andrew Leppla, Jorge Olmos, Jaideep Lamba

SMU Data Science Review

Non-Fungible Tokens (NFTs) enable ownership and transfer of digital assets using blockchain technology. As a relatively new financial asset class, NFTs lack robust oversight and regulations. These conditions create an environment that is susceptible to fraudulent activity and market manipulation schemes. This study examines the buyer-seller network transactional data from some of the most popular NFT marketplaces (e.g., AtomicHub, OpenSea) to identify and predict fraudulent activity. To accomplish this goal multiple features such as price, volume, and network metrics were extracted from NFT transactional data. These were fed into a Multiple-Scale Convolutional Neural Network that predicts suspected fraudulent activity based …

Self-Learning Algorithms For Intrusion Detection And Prevention Systems (Idps), Juan E. Nunez, Roger W. Tchegui Donfack, Rohit Rohit, Hayley Horn

SMU Data Science Review

Today, there is an increased risk to data privacy and information security due to cyberattacks that compromise data reliability and accessibility. New machine learning models are needed to detect and prevent these cyberattacks. One application of these models is cybersecurity threat detection and prevention systems that can create a baseline of a network's traffic patterns to detect anomalies without needing pre-labeled data; thus, enabling the identification of abnormal network events as threats. This research explored algorithms that can help automate anomaly detection on an enterprise network using Canadian Institute for Cybersecurity data. This study demonstrates that Neural Networks with Bayesian …

Federated Learning Framework Integrating Refined Cnn And Deep Regression Forests, Daniel Nolte, Omid Bazgir, Souparno Ghosh, Ranadip Pal

Department of Statistics: Faculty Publications

Predictive learning from medical data incurs additional challenge due to concerns over privacy and security of personal data. Federated learning, intentionally structured to preserve high level of privacy, is emerging to be an attractive way to generate cross-silo predictions in medical scenarios. However, the impact of severe population-level heterogeneity on federated learners is not well explored. In this article, we propose a methodology to detect presence of population heterogeneity in federated settings and propose a solution to handle such heterogeneity by developing a federated version of Deep Regression Forests. Additionally, we demonstrate that the recently conceptualized REpresentation of Features as …

Federated Learning Framework Integrating Refined Cnn And Deep Regression Forests, Daniel Nolte, Omid Bazgir, Souparno Ghosh, Ranadip Pal

Department of Statistics: Faculty Publications

Predictive learning from medical data incurs additional challenge due to concerns over privacy and security of personal data. Federated learning, intentionally structured to preserve high level of privacy, is emerging to be an attractive way to generate cross-silo predictions in medical scenarios. However, the impact of severe population-level heterogeneity on federated learners is not well explored. In this article, we propose a methodology to detect presence of population heterogeneity in federated settings and propose a solution to handle such heterogeneity by developing a federated version of Deep Regression Forests. Additionally, we demonstrate that the recently conceptualized REpresentation of Features as …

Fuzzy Kc Clustering Imputation For Missing Not At Random Data, Markku A. Malmi Jr.

USF Tampa Graduate Theses and Dissertations

Research has a variety of difficulties, especially when involving human subjects, and one of the most prevalent is the issue of missing data. Missing data will always be present in research due to the fact there is no perfect method for collecting data and protecting against human error or mechanical failure. This requires researchers to be able to mitigate the problems that come along with missing data; reduction in power of an analysis and bias introduced by the missing pattern. This research investigated a non-parametric method using a nested approach of fuzzy K-Modes and fuzzy C-Means clustering to impute missing …

On Characterization Of The Exponential Distribution Via Hypoexponential Distributions, George Yanev

School of Mathematical & Statistical Sciences Faculty Publications

The sum of independent, but not necessary identically distributed, exponential random variables follows a hypoexponential distribution. We focus on a particular case when all but one rate parameters of the exponential variables are identical. This is known as exponentially modified Erlang distribution in molecular biology. We prove a characterization of the exponential distribution, which complements previous characterizations via hypoexponential distribution with all rates different from each other.