Thesiss

In this thesis, I explore in two essays how to augment thin historical purchase data with other sources of information using Bayesian and probabilistic machine learning frameworks to better infer customers' preferences and their future behavior. In the first essay, I posit that …

One of the major threats to the security of software systems is the occurrence of security vulnerabilities, which can potentially cause a variety of problems including, but not limited to, information loss, privilege escalation, data breach, and system failure. Software vulnerab…

Deep neural networks have fundamentally transformed the field of machine translation, and replaced statistical phrase-based approaches to serve translations to millions of users in production systems every day. Despite impressive progress in translation accuracy, improving decod…

In recent years, journalists and scholars alike have grown wary about the state of political discourse in the US. Numerous scholars have indicated that a major reason for the precipitous decline in civil discourse is the proliferation of rights talk, a rhetorical strategy predic…

Though the majority of the Antarctic continent is covered by ice, portions of Antarctica, mainly in the McMurdo Dry Valleys and the Transantarctic Mountains (TAM), are currently ice-free. The soils which have developed on these surfaces have been re-worked by the advance and ret…

Probabilistic models have a rich history in machine learning, offering a theoretical and practical framework for learning from observed data. Probabilistic models describe relationships between observed data and latent variables in terms of probability distributions. Practitione…

Protozoan zoonoses, such Chagas disease and leishmaniasis, remain endemic in large parts of the world, exacerbating social inequity and contributing heavily to the global burden of infectious disease. Novel protozoa species which have emerged from wildlife to humans in the recen…

Falls are a traumatic and possibly life-threatening incident experienced by an increasing population of elderly adults. Starting at the age of 65, the risk of falls starts to grow exponentially. Even if they are not fatal, they cause physical injuries such as broken bones, brain…

Alcohol is the third leading preventable cause of death in the United States. Alcohol Use Disorder (AUD) is also highly prevalent in the United States, with approximately 14.1 million adults and 414,000 adolescents being affected. It is, therefore, extremely important to underst…

The primary goal of this dissertation is to contribute to a predictive model of Spanish heritage grammars by applying recent advancements in machine learning to ferret out factors and covariates that give rise to the Spanish subjunctive. Given that “[p]redictive models would ent…

In economics, it is increasingly recognized that "culture matters". In this light, it is important to understand why culture changes and how it affects economic outcomes. In my dissertation, I examine the causes and effects of changing culture on economic development through a f…

Providing effective risk assessments of the likelihood of recurrent maltreatment is one of the key responsibilities of child welfare services in preventing subsequent maltreatment of a child who is a victim of abuse or neglect. This dissertation aimed to address foundational res…

Electronic health records (EHRs) are growing at an exponential rate, and predictive analysis of this data can help physicians determine a patient’s condition at an early stage. In order to perform predictive analytics, various factors need to be considered, including demographic…

Unmanned Aerial Vehicles (UAVs), commonly known as drones, is an emerging technology with a huge potential to transform our lives into smart and connected communities by providing efficient infrastructure solutions for a wide range of military, industrial and commercial applicat…

The next generation of neutrinoless double beta (0νββ) decay experiments and direct detection of dark matter require ultrapure detectors. The University of South Dakota is leading the development of ultra-pure germanium crystals. The project includes purification work through zo…

The wind design of tall buildings is moving towards the performance-based design (PBD) methodology due mainly to its advantages in clearly articulating stakeholder expectations, explicitly satisfying performance objectives and considerably reducing building life-cycle cost. The …

Always-on sound classification is a desirable but power-intensive function for a variety of emerging applications such as wearables and IoT devices. The hardware energy consumption of sound classifiers is typically driven by signal digitization, feature extraction, classificatio…

Intrusion detection systems (IDS) using machine learning is a next generation tool to strengthen the cyber security of networks. Such systems possess the potential to detect zero-day attacks, attacks that are unknown to researchers and are occurring for the first time in history…

Purpose: Gastric cancer is responsible for about 1 million incident cancer cases and 750,000 cancer deaths annually, half of which occur in East Asia. Survival is low because the disease is usually diagnosed at late stage when treatment is difficult. However, early stage disease…

An underexplored aspect of the abundant linguistic data available for computational analysis is the potential for conducting large-scale comparative analyses between different discourses as a more rigorous complement to existing methods. A significant obstacle in comparative stu…