Data Science Commons^™

Predictive Modeling Of Colorectal Cancer Risk: Leveraging Health, Demographic, And Socioeconomic Factors For Targeted Screening, Anish Bhandari, Shawn Deng, Michael Olheiser, Robert Slater

SMU Data Science Review

Colorectal cancer (CRC) remains a significant public health concern, affecting millions in the United States and worldwide. This study investigates the risk factors associated with CRC using data from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial and aims to develop predictive models to identify high-risk individuals for targeted screening and increased awareness. The dataset integrates CRC incidence data from the National Cancer Institute with socioeconomic indicators from U.S. Census Bureau, linked by zip code. We employ Logistic Regression and Neural Network models to predict CRC risk, incorporating health, demographic, and socio-economic features. While the results suggest that …

Context-Switch Attacks: Understanding And Mitigating The Threat To Llm Applications, Sydney Holder, Bivin Sadler

SMU Data Science Review

Large Language Models (LLMs) are transforming conversational AI, yet their dependence on prompt-supplied context exposes them to context-switch attacks that covertly steer dialogue toward sensitive or malicious ends. A 70 one-sided conversation transcript evaluation set was constructed spanning various fraudulent scenarios. Each transcript embeds adversarial patterns drawn while preserving natural conversational flow. We introduce a hybrid defense that pairs a BERT-based semantic-drift detector (cosine-similarity threshold = 0.70) with a curated keyword and hack-phrase scanner to counter these threats. In aggregate, the system delivered 100 % recall, intercepting every simulated phishing or data-harvesting attempt. The keyword layer achieved perfect precision, generating …

The Prevalence And Impact Of Discourse In Social Media Networks: The 2024 Presidential Election, Kyle Kuberski, Xavier R. Mojica, Gwonchan J. Yoon, Brad Klein, Bivin P. Sadler

SMU Data Science Review

Abstract. Social Media platforms serve as central hubs for global discourse, where political dialogue is widely shared and echoed. This exchange shapes civic participation and influences electoral outcomes, often with both intended and unintended consequences. In its inception, social media platforms served as message boards for the masses, yet manipulation and exploiting of systems via bot usage has made platforms susceptible to outside forces. Thus, false narratives and an artificial sense of consensus are endemically augmented. As for the 2016 and 2020 U.S. presidential elections, it is important to investigate the prevalence of evolved bot activity in both political and …

Molecular Quantum Particle Algorithm (Mqpa): Hybrid Quantum-Classical Learning For Molecular Property Prediction, Jessica T. Mcphaul, Bivin Sadler Ph.D.

SMU Data Science Review

Classical machine learning models and quantum kernel methods often struggle to capture quantum-coherent molecular features under the constraints of noisy intermediate-scale quantum (NISQ) hardware, limiting both predictive accuracy and scalability.

This paper introduces the Molecular Quantum Particle Algorithm (MQPA), a hybrid quantum–classical framework designed to achieve chemically accurate property prediction by integrating handcrafted molecular descriptors with parameterized quantum circuits. Molecular inputs, expressed as SMILES strings, are processed via RDKit and encoded through angle-based quantum gates with entangling layers in Qiskit [1]. Quantum parameters are optimized using simultaneous perturbation stochastic approximation (SPSA) [2], while classical regression layers leverage Adam [3] with …

Mapping Responsible Workflows For Geospatial Data Science: Developing The I-Guide Data Ethics Toolkit, Peter T. Darch, Kyra M. Abrams, Ivan Y M Kong

I-GUIDE Forum

AI workflows in geospatial data science offer significant societal benefits but raise ethical, transparency, and reproducibility challenges. Current ethical frameworks and tools are often hard to integrate into daily research practice. This paper introduces the I-GUIDE Data Ethics Toolkit (DET), a lightweight suite designed for users of the NSF-funded Institute for Geospatial Understanding through an Integrative Discovery Environment (I-GUIDE). Based on a longitudinal mixed-methods study, including surveys, interviews, and observations, we identified five design priorities: usability, anticipatory planning, distributed responsibility, comprehensive coverage, and policy compliance. We integrated existing AI and data research lifecycles into an eight-stage I-GUIDE Research Lifecycle, serving …

Machine Learning-Based Variance Analysis Of Brightness Temperature In Simulated Satellite Footprints, Chhaya R. Kulkarni, Nikki Prive, Vandana P. Janeja

I-GUIDE Forum

This study investigates the variance in brightness temperature (BT) within simulated satellite footprints for Observing System Simulation Experiments (OSSE), focusing specifically on Channels 5 and 11 of the Advanced Microwave Sounding Unit (AMSU-A). High-resolution atmospheric simulations from the DYnamics of the Atmospheric general circulation Modeled On Non-hydrostatic Domains (DYAMOND) dataset were utilized to generate brightness temperature data using the Python interface for the Community Radiative Transfer Model (PyCRTM). A computational design map incorporating Random Forest and Association Rule Mining was employed to identify and validate key atmospheric variables influencing BT variance. This ensemble approach facilitated a deeper understanding of atmospheric …

Topic Shift Detection And Triggering In Natural Dialogue Systems: A Lightweight Approach, Rohith Perumandla

College of Computing and Digital Media Dissertations

This research address a key challenge in dialogue system: enabling the proactive, human-like shifting using lightweight approaching using MobileBERT (~25M) model was proposed and fine-tuned for topic shift detection, augmented with liguistic featuers for for topic trigger detection. Despite its smaller size (~25M parameters), the MobileBERT-based system achieved competitive results (F1 = 74.16%,) compared to the much larger XLNet model (~110M parameters, F1 = 79.95%), while offering greater efficiency. The topic trigger module, combining MobileBERT with linguistic features, further demonstrated effective performance (F1 = 71.61%).

Early Diagnosis And Detection Of Skin Cancer Using Deep Neural Network Models And Feature Extraction From Pre-Trained Cnns, Sahil Khanna

Harrisburg University Dissertations and Theses

Skin cancer is one of the most common and lethal cancer types. While accurate diagnosis at an early stage is essential for skin cancer treatment it remains difficult to achieve in many regions due to lack of sufficient dermatologists and proper diagnostic equipment. Prior studies show Convolutional Neural Network (CNN) models excel at skin lesion classification and consistently achieve better results than standard diagnostic practices. However, the focus of many studies remains confined to image-based learning while neglecting useful patient metadata that could improve prediction accuracy. This research project created a specialized CNN model to classify skin lesions and evaluated …

A Comparative Study Of Neural Networks And Xgboost Models For Flight Time Prediction, Ioannis Paraschos, Taryn E. Trimble, Eshna Bhargava, Jake Klingler, Benjamin R. Nicolai

Beyond: Undergraduate Research Journal

Flight time prediction plays a crucial role in modern air travel, benefiting airlines and passengers alike. Accurate predictions enable airlines to optimize schedules, allocate resources effectively, and ensure passenger safety and satisfaction. In recent years, machine learning models, such as neural networks and XGBoost, have gained popularity for predicting flight times. This study aims to compare the performance of neural network and XGBoost models in predicting flight times, considering factors such as weather conditions, air traffic control, and aircraft performance. The results indicate that both models are effective, with XGBoost achieving slightly higher accuracy. However, neural networks offer advantages in …

Data Sonification: The Art Of Exploring Space Through Sound, Jessica Cotturone

Music: Student Scholarship & Creative Works

Data sonification is a cutting-edge, interdisciplinary field that involves the transformation of data into sound. As the amount of data collected in today’s world continues to rise exponentially, new collaborations between scientists, mathematicians, and musicians are exploring the potential applications of sonification. Turning data into sound can create new avenues for interpretation and can increase accessibility for visually-impaired populations. The sonic renderings produced by this practice exist on a spectrum from sound to music and vary in their levels of objectivity and expressivity. One of the major areas where data sonification is providing a new scientific and artistic perspective is …

Neat Vs. Scruffy: How Early Ai Researchers Classified Epistemic Cultures Of Knowledge Representation, Lindsay Poirier

Statistical and Data Sciences: Faculty Publications

Accounts of the history of symbolic artificial intelligence (AI) often categorize researchers and their programs through binary oppositions: logicist versus procedural, or alternatively, neat versus scruffy. These terms were often referenced throughout the 1970s and 1980s by AI researchers at institutions in the United States and Europe to document epistemic and esthetic differences in approaches to knowledge representation. In this article, I historicize and situate these binary oppositions, showing how researchers leveraged them in discourse to demarcate the beliefs and commitments underpinning certain AI approaches. I argue that the labels helped knowledge representation scholars identify, trouble, and grapple with the …

Keynote - Data? We Don't Have Time For Data: A Realistic Look At Law Enforcement Use Of And Need For Human Trafficking Data, Doug Gilmer Phd

SMU Human Trafficking Data Conference

Drawing on over 35 years of law enforcement experience (25 years with the Department of Homeland Security), Dr. Gilmer will speak from a government and law enforcement perspective on the need and use for human trafficking data. Some agencies and components of the U.S. government, and individual states, are heavily invested in collecting data to satisfy their reporting requirements. From a law enforcement perspective, however, big human trafficking data sets are rarely examined. Data science in law enforcement is a relatively new phenomenon, and most law enforcement officers do not have the time, resources, or background to collect or analyze …

Comparative Analysis Of Matrix Factorization And Neural Collaborative Filtering For Movie Recommendation Systems, Mahyar Alinejad

Data Science and Data Mining

This paper presents a comparative study of two recommendation system approaches for predicting movie ratings: Matrix Factorization with Stochastic Gradient Descent (SGD) optimization and Neural Collaborative Filtering (NCF) using Tensor Flow. The study aims to evaluate the effectiveness of these methods in recommending movies to users based on the MovieLens 100K dataset. The Matrix Factorization approach utilizes latent features to model user preferences and item characteristics, optimizing parameters through SGD. On the other hand, NCF integrates traditional collaborative filtering with neural networks to capture complex user-item interactions. Experimental results demonstrate the performance of both models in terms of Root Mean …

Temporal Modeling And Forecasting Of Blood Glucose Dynamics In Individuals With Diabetes Mellitus, Mj Ruff

University Honors Theses

People living with Diabetes Mellitus face significant health risks, including an increased likelihood of heart disease, stroke, and fluctuations in blood glucose levels. The unpredictable nature of glucose levels can lead to dangerous conditions such as ketoacidosis and hypoglycemia. This study employs advanced time series analysis tools to forecast the glucose levels for an individual diagnosed with Type 1 Diabetes Mellitus.

Detecting Physical Activity Using Wearable Sensor Data, Dipok Deb

Data Science and Data Mining

This study focuses on detecting physical activity using wearable sensor data, specifically distinguishing between walking and running. A dataset comprising accelerometer and gyroscope readings is used to train and evaluate various machine learning models, including logistic regression, random forest, k-nearest neighbors, naïve Bayes, and XGBoost. Extensive preprocessing, such as creating lag features and rolling statistics, is performed to enhance temporal data representation. The models are evaluated using metrics like accuracy, precision, recall, and F1 score. Incorporating lag and rolling features significantly improves model performance, with logistic regression achieving perfect scores across all metrics. These findings demonstrate the effectiveness of enhanced …

Financial Data Stewardship: A Librarian's Approach To Data Discovery And A Call For Coding As A Path To Data Comprehension And Verification, Lip Hwe Tee

Research Collection Library

This conference presentation calls to advocate and emphasise coding as a means to understand datasets, to be able to correctly interpret table structure and relevant fields to extract the correct data, to ensure accuracy and relevance.

A Decision Support System For Conference Session Selection Using Natural Language Processing, Tillman E. Erb

Master's Theses

Conference attendees are faced with selecting from hundreds to thousands of presentations and sessions in pursuit of new findings and methods relevant to their area of interest, an overwhelming amount of information from which to clearly make a decision. To address this, we developed a decision support system leveraging natural language processing (NLP) techniques such as semantic matching. By creating and matching embeddings of conference presentation abstracts and titles, the application provides improved query matching compared to keyword searching. We introduce Session Scout, a novel conference decision support system built upon a semantic retrieval framework. Session Scout is designed to …

Density-Based And Model-Based Clustering With Tidyclust In R, Brendan S. Callender

Master's Theses

Clustering is a fundamental technique in unsupervised learning that can be used to find hidden patterns and structures within unlabeled data. The tidyclust package in R provides a unified interface for applying various clustering techniques to data. This paper outlines the addition of density-based clustering with DBSCAN, and model-based clustering using Gaussian mixture models (GMMs) to the tidyclust package. DBSCAN can be performed using the db_clust() function and makes use of the dbscan package implementation as its engine. GMMs can be fit using the gm_clust() function which makes use of the mclust package implementation. This paper highlights the changes made …

Topodino: Self-Supervised Topological Representation Learning For Neuronal Morphologies, Yasser Binbisher

Master's Theses

Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization …

Frequent Itemset Mining With Tidyclust In R, Andrew D. Kerr

Master's Theses

Unsupervised learning is closely associated with clustering, however other methods fall under this umbrella such as data mining. In R, the tidyclust package provides a unified interface for clustering models, yet lacks support for data mining. This thesis addresses this gap by introducing the Apriori and ECLAT algorithms into tidyclust, with a focus on frequent itemset mining. Unlike traditional clustering models, frequent itemsets produce groupings of column variables, rather than cluster labels or partitions of observations. To address this, a novel clustering approach is proposed: items (columns) are grouped based on their ”dominant” frequent itemset. A key contribution is a …