Digital Commons Network^™

Improving Rail System Signaling Efficiency Through Ai-Based Driving Profile Generation: A Comparative Performance Analysis, Mehmet Taci̇ddi̇n Akçay, Abdurrahi̇m Akgündoğdu

Turkish Journal of Electrical Engineering and Computer Sciences

In this study, a dataset comprising 3600 discrete operational snapshots (rather than continuous time-series data) derived from real-field operations is used to obtain a high-accuracy driving profile equation using a second-degree Polynomial Regression method. This equation demonstrates the model’s interpretability. The performance metrics obtained with the second-degree polynomial regression model’s equation are as follows: a coefficient of determination (R2) of 0.84, a Pearson Correlation Coefficient of 0.91, and an RMSE of 11.13. These results indicate the effectiveness of artificial intelligence-based approaches in improving the efficiency of the railway signaling system. The same dataset is also utilized with other machine learning …

Brain Tumor Segmentation Based On Convolutional Neural Network And Integration Of Deep And Shallow Features, Rula Sami Aleesa

Baghdad Science Journal

Segmentation of Brain tumors refers to one of the most challenging problems in analyzing a medical image. To establish accurate brain tumor region delineation, brain tumor segmentation is used. Feature extraction refers to one of the basic steps in the processing of an image that aids classification. Various features’ kinds are extracted from MRI images. The traditional classifiers of Machine learning need hand-crafted features that are time-consuming and susceptible to errors made by humans. Against this, Deep learning is too powerful in terms of feature extraction and has already been extensively applied for the aims of classification. The presented technique …

High-Performance Deep Learning Techniques For Plant Disease Detection: Performance Analysis, Validation, And Applications, Doaa El-Shahat, Ahmed Elmasry

Sustainable Machine Intelligence Journal

The early detection of plant diseases is an indispensable task to improve crop yields and production quality. Crop disease observations by experienced pathologists are difficult and might take a long time. Therefore, deep learning (DL) techniques have been utilized to present an automated detection technique that could accurately and timely detect plant diseases. Several DL models in the literature were proposed, but no paper conducted a comparative study between those models to determine which of them was the best alternative for this task. Therefore, twenty-one DL models are compared in this review paper to show which of them could achieve …

Rf-Et-Ann: Hybrid Machine Learning Model For Forecasting Short-Term Photovoltaic Power Production, Walid Abdullah, Ahmed Ismail Ebada, Mohamed Abouhawwash

Sustainable Machine Intelligence Journal

Accurately forecasting photovoltaic (PV) power generation is a challenging problem due to the non-linear and highly variable characteristics of solar data, which is strongly influenced by several interdependent factors, such as environmental conditions, system characteristics, and technical aspects. In this study, a newly proposed machine learning (ML) technique based on integrating random forest (RF), extra trees (ET), and artificial neural networks (ANN) is presented to tackle this problem with better predictive accuracy, ensuring that solar energy is used more consistently, effectively, and economically; this model is dubbed RF-ET-ANN for short. Hybridization of these three models will help capture non-linear aspects …

Unravelling Impact Of Comorbidities On Mortality Risks In Ckd Patients During The Covid-19 Pandemic: An Explainable Ai-Driven Study, Zeinab Abdollahi, Lin Huo, Donald Fraser, Shang-Ming Zhou

School of Nursing and Midwifery

ObjectivesThe chronic kidney disease (CKD) patients were at high risk for severe clinical complications during the COVID-19 pandemic. Our objectives were to evaluate comorbidity prevalence; predict mortality risks for CKD patients during the pandemic; assess how various health factors interact to influence mortality; and provide insights for targeted prevention strategies.MethodWe analysed data from 186,396 CKD patients in Mexico during the entire pandemic (Jan 2020- May 2023). Explainable artificial intelligence (XAI) methods with extreme gradient boosting (XGBoost) models and Shapley Additive Explanations (SHAP) were developed to predict mortality for CKD patients with model interpretations. Different metrics were used to comprehensively evaluate …

Utilizing Machine Learning Techniques For Computer-Aided Covid-19 Screening Based On Clinical Data, Honglun Xu, Andrews T. Anum, Michael Pokojovy, Sreenath Chalil Madathil, Yuxin Wen, Md. Fashiar Rahman, Tzu-Liang Bill Tseng, Scott Moen, Eric Walser

Engineering Faculty Articles and Research

The COVID-19 pandemic has highlighted the importance of rapid clinical decision-making to facilitate the efficient usage of healthcare resources. Over the past decade, machine learning (ML) has caused a tectonic shift in healthcare, empowering data-driven prediction and decision-making. Recent research demonstrates how ML was used to respond to the COVID-19 pandemic. This paper puts forth new computer-aided COVID-19 disease screening techniques using six classes of ML algorithms (including penalized logistic regression, random forest, artificial neural networks, and support vector machines) and evaluates their performance when applied to a real-world clinical dataset containing patients’ demographic information and vital indices (such as …

A Contextual Attention-Based Transformer Model For Enhanced Hate Speech Detection On Twitter, Mira Mansour, Nancy Akoum, Seifedine Kadry

Iraqi Journal for Computer Science and Mathematics

Hate speech on social media poses significant societal challenges, necessitating accurate and context-sensitive automated detection. Traditional machine learning (ML) models typically rely on lexical or superficial features, limiting their ability to capture nuanced or contextually ambiguous expressions of hate speech. Recent transformer-based methods (e.g., RoBERTa) provide improved contextual understanding but often lack explicit mechanisms guiding the model’s attention to critical semantic tokens, thereby reducing interpretability and sensitivity to nuanced linguistic contexts. This paper introduces a novel contextual attention-guided transformer model that explicitly incorporates lexicon-guided attention supervision into RoBERTa fine-tuning, significantly enhancing semantic precision in hate speech detection on Twitter. Evaluations …

Optimizing Proaftn Classifier With Ant Colony Algorithm: Enhanced Diabetes Detection Benchmarking, Feras Al-Obeidat

All Works

The increasing global prevalence of diabetes highlights the need for accurate diagnostic tools to improve early detection and effective treatment planning. Traditional classification models often struggle to achieve optimal performance due to limitations in parameter tuning and adaptability to complex datasets. To address these limitations, this article introduces PROAnt, an innovative learning approach designed to enhance the robustness and efficiency of the PROAFTN multicriteria classification method. PROAnt leverages the computational power of ant colony optimization (ACO) to dynamically fine-tune and optimize the key parameters, such as intervals and weights, at the core of the PROAFTN classification process. This learning methodology …

A Visualization-Supported, Hierarchical, Action-Learning Model For Driving Behavior In A V2x Environment, Xuantong Wang, Jing Li, Jecca Bowen

Geography and the Environment: Faculty Scholarship

Understanding human driving decisions is crucial for intelligent transportation research. Most existing studies focus on individual vehicles in limited contexts, which restricts broader applicability of results. Leveraging Vehicle-to-Everything (V2X) infrastructure, this study introduces a machine learning framework to model driving actions and detect outliers across diverse environments. This approach features a semantically enabled clustering method that groups similar driving behaviors based on speed and actions. It also adds a time-series learning model to identify typical driving behaviors across various contexts, thereby enabling detection of abnormal driving actions. A suite of visual tools has been developed to help interpret driving patterns, …

Timing, Orbital Pacing, And Provenance Of Late Paleogene Loess In The Western United States, Xiangwei Guo

Earth & Environmental Sciences Dissertations - Archive

This project examines the timing of loess emergence, the nature of the fluvial–eolian transition, orbital forcing on dust accumulation, and sediment recycling in the late Paleogene White River Formation (Group) of Wyoming and Nebraska, western United States. By integrating traditional sedimentology with machine learning–enhanced grain-size analysis, this research shows that loess accumulation at Flagstaff Rim in central Wyoming began during active fluvial deposition at approximately 35.8 Ma, indicating a gradual transition from fluvial to eolian conditions. Machine learning complements field sedimentology and granulometric interpretation while enabling the development of new, testable hypotheses. Extending this framework eastward to Toadstool Geologic Park …

Utilizing Machine Learning Techniques For Computer-Aided Covid-19 Screening Based On Clinical Data, Honglun Xu, Andrews T. Anum, Michael Pokojovy, Sreenath Chalil Madathil, Yuxin Wen, Md Fashiar Rahman, Tzu-Liang (Bill) Tseng, Scott Moen, Eric Walser

Mathematics & Statistics Faculty Publications

The COVID-19 pandemic has highlighted the importance of rapid clinical decision-making to facilitate the efficient usage of healthcare resources. Over the past decade, machine learning (ML) has caused a tectonic shift in healthcare, empowering data-driven prediction and decision-making. Recent research demonstrates how ML was used to respond to the COVID-19 pandemic. This paper puts forth new computer-aided COVID-19 disease screening techniques using six classes of ML algorithms (including penalized logistic regression, random forest, artificial neural networks, and support vector machines) and evaluates their performance when applied to a real-world clinical dataset containing patients’ demographic information and vital indices (such as …

Ai For Life Sciences: From Geometric Protein Modeling To Multimodal Drug Design, Feng Jiang

Computer Science and Engineering Dissertations

Predicting biomolecular interactions, from immune recognition to drug–target binding, is a central problem in the life sciences and computational drug discovery. Deep learning has advanced this area, yet three challenges persist: the topology of large, highly imbalanced interaction networks; structural noise in computationally predicted protein models; and the integration of multimodal information such as functional text and taxonomic annotations. This dissertation develops a coherent set of models spanning immune complex prediction and small-molecule drug design: graph learning that addresses network topology and severe class imbalance; a noise-tolerant method that fuses predicted structures with evolutionary sequence features; and multimodal representation learning …

Comparative Machine Learning Models For Disease Risk Prediction, Mercy Mawusi Agbley

Theses, Dissertations and Capstones

Accurate prediction of disease outcomes is crucial for improving clinical decision-making and enabling early intervention. This study compares the performance of various statistical and machine learning models for clinical risk prediction using two healthcare datasets: diabetic retinopathy and heart disease. The models assessed include Logistic Regression, LASSO, k-Nearest Neighbors (KNN), Support Vector Machines (SVM), Neural Networks, Random Forests, Gradient Boosting Machines (GBM), and a stacked ensemble model. Prior to modeling, datasets were split into train and test sets. Standardization was applied to numeric features whilst categorical features were one-hot encoded. These transformations were later applied to the test set. Principal …

A Hybrid Response Surface Methodology And Machine Learning Framework For Quantifying Effects Of Physicochemical Parameters On Pfas Distribution, Harsh V. Patel, Jazmin Green, Hyoshin Park, Stephanie Luster-Teasley Pass, Renzun Zhao

Engineering Management & Systems Engineering Faculty Publications

Predicting PFAS adsorption across diverse adsorbents and environmental matrices remains challenging because adsorbent physicochemical properties, PFAS molecular descriptors, and operational conditions simultaneously influence adsorption. This study develops and evaluates a unified hybrid modeling framework that integrates Response Surface Model (RSM) with machine-learning algorithms to quantify how six key variables, surface area, Log K_ow, pH_pzc, pK_a, log dose, and log-initial concentration, affect PFAS distribution coefficients (Log K_d). A data set of more than 1000 adsorption observations spanning 15 PFAS compounds, multiple adsorbent types, and a broad operational range was compiled and preprocessed using …

Optimizing Machine Learning Algorithms Through Hyper-Parameter Tuning For Accurate Rice Yield Forecasting In North-East, Nigeria. International Journal Of Agricultural And Statistical Sciences, Ezra Daniel Dzarma, Guy Degla, Theophile Komlan Dagba, Nyor Ngutor

All Rsif Scholars' Publications

Accurate prediction of rice yield is crucial for strengthening food security and improving agricultural decisionmaking in North-East Nigeria, where production systems are constrained by fluctuating input use and environmental variability. This study applies four machine learning algorithms: Random Forest (RF), Extreme Gradient Boosting (XGB), Support Vector Regression (SVR), and K-Nearest Neighbours (KNN)to model rice yield using five primary farm inputs: Labour (B), Fertilizer (F), Herbicides (H), Seeds (S), and land (L)area. All models were optimized through hyper-parameter tuning to ensure reliable performance. The results show that SVM and XGB produced the strongest predictive accuracy, with RF achieving an RMSE of …

Serum Biomarker Trajectory Clusters Predict Functional Outcome And Quality Of Life For Traumatic Brain Injury, Thanh Son Do, Chantal Carnes, Zhihui Yang, Firas Kobeissy, Hamad Yadikar, Gayla R. Olbricht, Olli Tenovuo, Jussi P. Posti, Ewout W. Steyerberg, Lindsay Wilson, Nicole Von Steinbüchel, Endre Czeiter, Andras Buki, David K. Menon

Mathematics and Statistics Faculty Research & Creative Works

Serum brain-enriched biomarkers are increasingly employed in the clinical evaluation of traumatic brain injury (TBI) to assist with triage, neuroimaging decisions, and prognostication. However, the potential of temporal biomarker trajectories to inform disease monitoring and long-term outcomes remains underexplored. We aim to identify distinct biomarker trajectory (TRAJ) profiles in traumatic brain injury patients and to examine their associations with long-term clinical outcomes. The study included 373, CT-positive Intensive Care Unit (ICU) traumatic brain injury patients (256 with initial Glasgow Coma Scale 3–12) from the Collaborative European Neurotrauma Effectiveness Research in TBI (CENTER-TBI) core study who had at least two serum …

Ordered Mini-Batch Training For Differentially Private And Encrypted Logistic Regression, Ryan Leone

Theses, Dissertations and Culminating Projects

Logistic regression has found extensive use as a supervised machine learning algorithm due to its simplicity and efficiency in binary and multivariate classification tasks. As data sharing grows across connected devices, safeguarding sensitive personal and industrial information is of increased importance. Privacy-preserving machine learning techniques such as differential privacy and homomorphic encryption offer mathematically rigorous security guarantees, but introduce difficult accuracy, privacy loss, and computational overhead issues. This thesis investigates PPML for logistic regression through a collaborative mini-batch training framework. I propose and implement an ordered mini-batch strategy, compare it to standard shuffled methods, then integrate differential privacy noise injection …

Student Retention In Music Programs: An Analysis Of Relationships Between Selected Student Demographics And Continued Enrollment In Music Programs With The Presence Of School-Provided Supports, Jessica Wiese

Theses, Dissertations and Capstones

Music education provides significant cognitive, social, and academic benefits; however, elective music programs often face substantial annual attrition. While existing literature has documented material barriers (e.g., instrument costs, transportation, scheduling conflicts), this study investigates the persistence of attrition within a medium-sized suburban district on Long Island that has systematically mitigated these obstacles. The district provides free instruments, flexible scheduling, and dedicated transportation for all students. The purpose of this mixed-methods study was to determine whether demographic disparities in music enrollment remain predictive of withdrawal in such a post-barrier environment.

The research used a two-phase explanatory sequential design. In the first …

Advances And Challenges In Digitally Connected Point-Of-Care Biosensing, Abdellatif Ait Lahcen, Jegan Rajendran, Gymama Slaughter

Center for Bioelectronics Publications

Point-of-care (POC) biosensors are undergoing a paradigm shift from isolated diagnostic tools to digitally connected, intelligent platforms that enable continuous and decentralized healthcare delivery. This review critically examines recent advances in wearable, implantable, and portable biosensors, highlighting how integration with wireless communication, the Internet of Medical Things (IoMT), and artificial intelligence is transforming their functionality and clinical utility. Particular attention is given to innovations such as smartphone-enabled interfaces, cloud-based analytics, and machine learning-assisted analysis, which collectively enhance sensitivity, specificity, and user accessibility across diverse healthcare settings, from personalized home monitoring and bedside diagnostics to deployment in resource-limited regions. The review …

Training Set Augmentation And Biology-Aware Harmonization Improve Radiomic Models For Lung Cancer Prediction In Indeterminate Nodules, Claire Huchthausen, Menglin Shi, Gabriel De Sousa, James Larner, Einsley Janowski, Jonathan Colen, Krishni Wijesooriya

Data Science Faculty Publications

CT radiomics-based machine learning has potential to predict lung cancer in pulmonary nodules (PNs) earlier than standard-of-care methods. Low malignancy rates in early-development PNs and variable image acquisition hinder development of radiomic models for diagnosing these PNs. To address these challenges, we augmented training using later-development PNs and harmonized for acquisition effects. We examine early-development benign and malignant PNs (n = 106) below the sensitivity of standard-of-care diagnosis. Classifiers predicting malignancy performed near chance when trained on ComBat-harmonized radiomic features from only early-development PNs. We then augmented training with later-development benign and malignant PNs (n = 225). We evaluated whether …

Stall Detection In Hydraulic Excavator Operations Using Heuristics And Machine Learning: A Case Study, Mateo Fernando Montenegro Defaz, Kwame Awuah-Offei, Yixiang Gao

Mining Engineering Faculty Research & Creative Works

This work aims to develop a reliable algorithm for stall detection during excavator digging by analyzing key operational variables such as velocity and angular displacements from machine monitoring data. The work develops and validates a heuristic algorithm to detect stalling events and trains a support vector machine classification algorithm to distinguish between "normal" digging cycles and cycles with stalling. This work is a novel attempt at using a classification algorithm to categorize digging cycles into normal and those with stalling events based on machine monitoring data alone. The developed classification algorithm achieved a sensitivity of 100%, indicating it correctly identified …

Data-Driven Climate Damage Functions For Capital Formation: Estimating The Climate Penalty Using Maching Learning, Pramudya Wicaksono

All Graduate Theses, Dissertations, and Other Capstone Projects

Traditional integrated assessment models assume parametric climate damage functions that may miss nonlinearities, heterogeneity, and dynamic effects on investment. This thesis develops a data-driven climate damage function for capital formation by estimating the predictive relationship between climate conditions and future gross fixed capital formation (% GDP) across 125 countries over 1982–2019. We construct a panel dataset by combining daily ERA5 climate reanalysis data (accessed via the Copernicus Climate Data Store API and aggregated to yearly country-level variables including temperature anomalies, extreme heat days, frost days, precipitation, and solar radiation) with economic indicators from the World Bank World Development Indicators and …

Array Signal Processing And Machine Learning In 5g/6g Networks, Roopesh Kumar Polaganga

Electrical Engineering Dissertations - Archive

This dissertation investigates advanced methodologies in Array Signal Processing (ASP) and Machine Learning (ML) to enhance the performance, efficiency, and intelligence of next-generation wireless networks, with a primary focus on 5G and emerging 6G systems. As wireless networks face rapid traffic growth, increasingly heterogeneous service requirements, and more complex propagation environments, conventional design and optimization approaches become insufficient to meet evolving demands in reliability, capacity, spectral efficiency, and energy efficiency. On the network intelligence side, this work develops data-driven frameworks for causal discovery, scheduler enhancement, session-duration prediction, and Radio Resource Control (RRC) state optimization using real-world telecommunication network data. On …

Neighborhood Embeddings And Scalable Learning For Optimal Transport And Unbalanced Optimal Transport, Muhammad S. Rana

Mathematics Dissertations - Archive

Dimensionality reduction techniques are developed from the assumption that high-dimensional data often arises from low-dimensional structures embedded into the high-dimensional ambient space. Classical dimensionality reduction methods rely on the Euclidean distance, which may fail to capture the geometric structures of the datasets. This dissertation includes alternative metrics for dimensionality reduction techniques and challenges in applying these techniques.

First, we investigate the Wasserstein distance based neighbor embeddings for dimensionality reduction methods and compare the classification and clustering performance with the classical Euclidean based methods. The Wasserstein distance models the data as probability distributions and compares two distributions applying optimal transport (OT) …

Understanding Phishing Susceptibility Through Expert Consensus Using Digital Marketing Parallels And A Machine Learning-Based Implementation, Mansoor Ahmad

All Graduate Theses, Dissertations, and Other Capstone Projects

Phishing remains one of the most effective attack vectors for gaining unauthorized access to organizational systems, yet defenders often lack systematic methods to assess their exposure before an attack. This study develops a framework that uses machine learning to encode the collective expertise of cybersecurity practitioners into a portable phishing susceptibility assessment tool, with the goal to help security teams proactively identify patterns, prioritize awareness training, and strengthen detection controls. The study surveyed 27 practitioners with extensive experience in social engineering, red teaming, penetration testing, and threat analysis to identify which factors most influence phishing susceptibility. Practitioners provided quantitative ratings …

Analysis Of Surrogate Models At Multiple Levels For Neural Acceleration Of Hpc Applications, Bibek Panthi

Theses

Due to rise of machine learning workloads, high performance computing platforms have increased GPU resources. Consequently, use of surrogate models, to utilize GPU compute and speedup scientific applications, is on the rise. Most of the time, the whole program is replaced with a surrogate model. In this work a shock simulation program (LULESH) was taken and three functions of varying complexity were replaced with small sized neural network as surrogate models. The speedup and error in output of whole program was analyzed. We found that for the smallest function, trained model lead to speedup of overall program by 40%, maintained …

Development Of Multimodal Measurements And Analysis For Early Detection Of Alzheimer’S Disease, Fiza Saeed

Bioengineering Dissertations

Alzheimer's disease (AD) is the leading cause of dementia, and existing diagnostic methods such as PET scans, cerebrospinal fluid sampling and biomarker quantification, and gene sequencing are all either invasive, costly, or not sensitive enough for early detection. This dissertation introduces three different studies that develop a novel multimodal, non-invasive approach to diagnosing AD at its early stages by combining broad band near infrared spectroscopy (bbNIRS) and electroencephalography (EEG) technologies.

The first study showed cerebrovascular-cerebrospinal fluid coupling (CBV-CSF), which is measured by using 2-channel bbNIRS as an indicator of brain aging and early AD. Linear correlations between total blood (Δ[HbT]) …

Machine Learning-Based Lifetime Prediction Of Lithium Batteries: A Comparative Assessment For Electric Vehicle Applications, Abdelilah Hammou, Raffaele Petrone, Demba Diallo, Boubekeur Tala-Ighil, Philippe Makany Boussiengue, Hicham Chaoui, Hamid Gualous

Electrical & Computer Engineering Faculty Publications

This paper evaluates and compares four data-driven methods (Gaussian Process Regression (GPR), echo state network (ESN), gated recurrent unit (GRU), and long short-term memory (LSTM)) for lithium-ion capacity prognostics adapted to electric vehicle conditions. This comparison aims to find the most efficient prognosis method considering two constraints: the limitation of computational power and the unavailability of on-board capacity measurement that requires full charge and discharge conditions. The machine learning models are trained using capacity values estimated under vehicle conditions. The ageing data is collected from cycling tests of two battery chemistries, Lithium Fer Phosphate (LFP) and Nickel Manganese Cobalt (NMC), …

Interpretable Battery Soh Prediction: A Comparative Interpretability Framework For Multi-Architecture Ml Models, Shafiyee Islam, Gon Namkoong

Electrical & Computer Engineering Faculty Publications

This work introduces a unified interpretability-efficiency framework for lithium-ion battery state of health (SOH) prediction using hybrid deep learning architectures. We comparatively analyze four hybrid models: CNN LSTM MultiHead, CNN Feature Extractor LSTM, DNN LSTM, and DNN BiLSTM to disentangle how network topology, feature composition, and computational design influence both predictive fidelity and physical interpretability. By integrating Monte Carlo Shapley (MC Shapley), background occlusion SHAP (BoSHAP), and ablation analysis, we quantify the contribution and robustness of five electrochemical feature groups: time, capacity, voltage, dQ/dV and peaks of dQ/dV from NASA battery dataset. The results reveal a consistent dominance of differential …

Machine Learning Classification Of Prostate Cancer Genomic Sequences Using K-Mer And Sequence-Derived Features, Kuldeep Rawat, Hirendra Nath Banerjee, Jamie Noble, Saa Naudia Deloatch, Satyendra Banerjee, Sachin Shetty, Soumya Banerjee

VMASC Publications

Prostate cancer disproportionately impacts African American men, who experience significantly higher mortality rates and earlier disease onset than other populations. Current diagnostic approaches, including prostate-specific antigen testing and biopsy, lack sufficient specificity and sensitivity, underscoring the need for accurate, molecular-level classification tools. This paper presents a machine learning framework for binary classification of genomic DNA sequences as cancerous or healthy. A dataset of 1684 FASTA-formatted sequences obtained from the National Library of Medicine - GenBank was analyzed, with 1662 sequences retained after quality control filtering. Feature engineering yielded 67 attributes, including GC content, Shannon entropy, sequence length, and trinucleotide k-mer …