Digital Commons @ Florida Atlantic
Not a member yet
400 research outputs found
Sort by
GEOSPATIAL METHODS FOR INTERPRETING AND MANAGING COASTAL LANDSCAPE DYNAMICS
Full text linkCoastal landscapes are dynamic interfaces shaped by natural forces and human activity, requiring precise, multiscale monitoring for effective conservation. This research investigates some of the coastal monitoring challenges related to shoreline erosion, terrain modeling, and mangrove species classification at the Jupiter Inlet Lighthouse Outstanding Natural Area (ONA), a 120-acre coastal preserve in southeast Florida. The first study applies conditional entropy and partial correlation to assess the relationship between shoreline retreat and environmental variables using UAS-derived shoreline data collected from 2017 to 2023. The second study enhances Digital Terrain Model (DTM) accuracy, a critical factor in erosion assessments, by applying RandLA-Net, a deep learning model, to classify points derived from UAS-SfM point clouds. The third study uses UAV-based hyperspectral imagery and classification techniques to map mangrove species, producing a distribution map that supports future spatiotemporal monitoring and guides replanting strategies in erosion-prone areas. Collectively, these studies integrate geospatial technologies and statistical modeling to advance data-driven coastal monitoring strategies, informing shoreline stabilization and vegetation management under evolving climatic and anthropogenic pressures
A QUALITY IMPROVEMENT PROJECT TO STANDARDIZE SEXUAL DYSFUNCTION SCREENING IN PRIMARY CARE
No full textSexual health is an essential aspect of overall well-being but is often overlooked in primary care due to provider discomfort, time constraints, and lack of standardized screening protocols. This quality improvement (QI) project aimed to implement a consistent approach to screening for sexual dysfunction using the Arizona Sexual Experiences Scale (ASEX), a validated five-item tool assessing sex drive, arousal, orgasm, and satisfaction. Guided by Swanson’s Theory of Caring and Pender’s Health Promotion Model, the project emphasized compassionate, patient-centered care and the promotion of health-enhancing behaviors.
Using the Plan-Do-Study-Act (PDSA) framework, the ASEX tool was incorporated into annual wellness visits following provider training. Seven of ten providers attended the training; two reported prior awareness of the tool. Post-training surveys indicated strong willingness to implement ASEX (mean = 4/5). However, only two providers implemented the screening as part of their routine. Over three months, 30 patients were screened, with ASEX scores ranging from 5 to 20. Female participants reported slightly higher mean scores (M = 13.83) than males (M = 11.39), though no statistically significant gender or age differences were identified (p \u3e .05). Two patients screened positive and received appropriate evaluation and referrals.
Findings highlight that sexual dysfunction affects adults across demographics, reinforcing the need for routine screening in primary care. Sustaining ASEX integration through education, chart audits, and EHR prompts can improve early detection, enhance communication, and promote holistic patient care. Embedding caring practices into sexual health assessment, can create a supportive environment and promotes overall well-being. Future DNP initiatives should examine long-term outcomes and provider education strategies to strengthen sexual health assessment practices
ENHANCING TD SCREENING WITH THE AIMS SCALE: A QUALITY IMPROVEMENT PROJECT
No full textTardive dyskinesia (TD) is a persistent, often irreversible movement disorder associated with long-term use of dopamine receptor–blocking agents. Early identification is critical to prevent progression and minimize functional impairment; however, routine TD screening remains inconsistent in many clinical settings despite the availability of the Abnormal Involuntary Movement Scale (AIMS), the evidence-based gold-standard tool. This quality improvement project aimed to enhance TD screening in an outpatient psychiatric setting through an educational intervention designed to improve provider knowledge of risk factors, diagnostic criteria, and screening guidelines. Data were collected retrospectively and prospectively from patient encounters, with 33 pre-intervention and 35 post-interventions. AIMS completion increased from 54.5% pre-intervention to 71.4% post-intervention. Three patients screened positive for TD following the intervention, whereas no cases were identified pre-intervention. All patients who screened positive received appropriate treatment, including initiation of VMAT2 inhibitors (Austedo or Ingrezza) and adjustments to antipsychotic medications. Although all the goals were not met, clinically meaningful improvements in screening and identification of TD were observed, demonstrating the value of targeted educational interventions in promoting early detection and treatment
HYPERSPECTRAL IMAGE CHARACTERIZATION OF DEGRADED FIELD RAILROAD BALLAST UNDER VARYING FOULING AND WATER CONTENTS: POTENTIALS FOR REUSE AND SUSTAINABILITY
No full textRailroad ballast performance is affected by fouling materials that contain plastic and granular properties. The characterization of fouling materials are important factors in assessing track performance and maintenance requirements. This study presents the
hyperspectral reflectance characteristics of the degraded field railroad ballast under varying fouling contents (FCs) and water contents (WCs). The study involves experiments including material selections, preparation of railroad ballast samples, fouling materials, Hyperspectral Imaging (HSI) data acquisition and calibration, methodology for regression using Gaussian Process Regression and Artificial Neural Network (ANN) Regression, discussions, and conclusions. Gaussian and ANN Regression models are applied to generate the actual and predicted reflectances at VNIR wavelength range (400-1000 nm) to the degraded field ballast samples under varying FCs with Moderately Clean (5% FC), Moderately Fouled (15% FC), and Fouled (25% FC) with varying WCs of 0%, 20%, 40%, 60% and 80%. The prediction generalizes more effectively and accounts for varying water and fouling scenarios. The study presents discussions on a prospective framework of reusing screened and recycled degraded ballast with fresh ballast in railroad infrastructure and sustainability. Blending of degraded ballast with virgin ballast will contribute to a sustainable and cost-effective strategy in railroad track maintenance
AI-POWERED PREDICTION OF CO2 DISSOLUTION IN SEAWATER FOR ENHANCED CARBON CAPTURE EFFICIENCY
No full textClimate change remains one of the most pressing global challenges, with carbon dioxide (CO2) emissions playing a crucial role in exacerbating global warming and ocean acidification. Among emerging mitigation strategies, carbon capture and storage (CCS) stands out as a promising solution. This dissertation explores a novel carbon capture approach utilizing seawater and nickel nanoparticles (NiNPs) stabilized by carboxymethylcellulose (CMC) to enhance CO2 dissolution. Unlike conventional amine-based methods, which are resource-intensive and reliant on freshwater, this approach offers a more sustainable and environmentally friendly alternative. To address the high cost and complexity of experimental CO2 capture studies, a predictive artificial intelligence (AI) framework was developed to estimate CO2 dissolution efficiency as a function of NiNP and CMC concentrations. Because the experimental dataset includes only 24 measurements used for AI algorithm development, a data augmentation technique - monotone spline interpolation - was employed to generate a synthetic dataset of 24,000 points, enabling robust AI model training and improved generalization. Two AI models, Adaptive Neuro-Fuzzy Inference System (ANFIS) and Gaussian Process Regression (GPR), specialized for the limited input datasets, were trained and validated using performance indicators such as the Mean Absolute Percentage Error () and the coefficient of determination (2). ANFIS achieved a of 0.0212 and an 2 of 0.9880, while GPR achieved a of 0.0015 and an 2 of 0.9999. A comparative analysis identified the GPR model as the most reliable, with an accuracy of 95.7 %. This work introduces a scalable, data-efficient methodology for predicting CO2 capture performance, supporting the development of more adaptive and sustainable carbon capture technologies
UNDERSTANDING THE IMPACTS OF CAREER SERVICES WITHIN GRADUATE PROGRAMS: SPORT MANAGEMENT AND MASTER OF BUSINESS ADMINISTRATION
No full textThis quantitative study evaluated career-related outcomes of career coach interventions among proactive graduate students in a college of business. It investigated how career services interventions influence key areas like employment, job growth, resume and cover letter development, job search skills, LinkedIn strategies, networking, salary negotiation, and job placement. The study was guided by three research questions: (a) do graduate students\u27 perceptions of satisfaction regarding career-related outcomes change after meeting with a career service professional, (b) what type of career-related outcomes do graduate students seek from career services within a college of business, and (c) do perceived career-related outcomes vary depending on the level of proactivity among graduate students who receive career coaching? The study used a survey, which included a Career Engagement Scale that captured students’ post-meeting perceptions (Hirschi et al., 2014). The study’s significance lies in addressing the limited research on graduate-level career services, offering insights for higher education practitioners to enhance career services with development needs, and ultimately improving graduate employability. The findings of this study conveyed that graduate students who met with career services professionals reported positive experiences and valued the support, but their level of proactiveness did not significantly influence their perceived career outcomes
DEEP LEARNING-BASED LUNG NODULE CLASSIFICATION USING CT SCANS
No full textLung cancer is still one of the predominant reason of cancer deaths around the world. It usually starts with small lung nodules that can be seen on CT scans. It is very important to find these nodules early and correctly classify them so that the diagnosis and treatment can be done quickly and effectively. Nevertheless, it is difficult to differentiate between benign and malignant nodules because of their comparable imaging appearance and the variation in manual interpretation among radiologists. To overcome these constraints, our study introduces a deep learning-based CNN model for the automated classification of lung nodules in CT images, aiming to facilitate early detection and alleviate the workload of radiologists. A total of 130 patient cases with corresponding XML annotations were obtained from the LIDC-IDRI dataset. Patient-wise splitting was performed to avoid data overlap, dividing cases into 70% training, 15% validation, and 15% testing sets. Using the first radiologist’s reading, only nodules with malignancy scores of 1-2 (benign) and 4-5 (malignant) were included, while ambiguous cases (score 3) were excluded. Around each valid nodule, 128×128-pixel image patches were extracted, converted to RGB, resized to 224×224 pixels. Manual augmentation (flips and +10° rotation) was applied to benign patches in training and validation sets to address class imbalance, yielding 948 total patches (720 training, 130 validation, and 98 testing). The proposed network was based on EfficientNetB0 (pretrained on ImageNet), serving as a fixed feature extractor, followed by a GlobalAveragePooling2D layer, dropout (0.2), and a final sigmoid-activated dense layer for binary classification. The model was trained on the training data with validation monitoring and evaluated on independent test data at both patch and patient levels. At the patch-level, it achieved 81% accuracy, 87% precision, 60% specificity, 88% recall, and an 87% F1-score. At the patient-level, the model achieved 73% accuracy, 67% precision, 86% recall, 75% F1score, and 63% specificity. Grad-CAM was used for interpretability to highlight the regions that influenced model predictions. Overall, the proposed EfficientB0-based method demonstrates the strong potential for binary lung nodule classification from CT patches. Given the dataset size and encouraging performance outcomes, this approach shows promise for future development of reliable AI-based tools for lung cancer diagnosis
DO ALGAL ENDOSYMBIONT COMMUNITIES IN CORAL OUTPLANTS CHANGE OVER TIME? A SPATIO-TEMPORAL ANALYSIS OF ALGAL SYMBIONTS IN A SOUTH FLORIDA CORAL RESTORATION EXPERIMENT
No full textIncreasingly frequent marine heatwaves and bleaching events threaten the persistence of corals on Florida’s Reefs. Active restoration through outplanting is being implemented across multiple species and regions, yet its effectiveness under rising thermal stress remains uncertain. We analyzed Symbiodiniaceae ITS2 sequence data from outplants of three coral species, Pseudodiploria clivosa, Montastraea cavernosa, and Orbicella faveolata, sampled before (2023) and after (2024) a mass bleaching event. Multivariate analyses revealed species-specific yet regionally divergent trends: southern restoration regions exhibited pronounced shifts toward thermotolerant Durusdinium, while northern regions remained comparatively stable. Although fine-scale ITS2 variation contributed to community heterogeneity, genus-level shifts more clearly reflected ecological responses to stress. These results underscore that algal symbiont restructuring is both species- and region-dependent, emphasizing the need to integrate regional thermal regimes and algal symbiont flexibility into restoration planning under climate change
NOVEL UNSUPERVISED FRAMEWORKS FOR AUTOMATED CLASS DISTRIBUTION ESTIMATION AND BINARY LABEL GENERATION FOR TABULAR DATA
No full textIn today’s landscape, copious amounts of unlabeled data continue to be generated. This data has the potential to accelerate machine learning research; however, supervised methods require labels and unsupervised methods often require expert fine-tuning to be reliable, both of which can impose significant cost. In addition to not requiring labels, another benefit of unsupervised learning is the protection of privacy since it does not require human annotation. In addition, class imbalance, where one class has significantly more instances, can complicate model training and reduce performance. Because of these challenges, automated unsupervised methods can offer a path forward to further machine learning research. The primary objective of this dissertation is to develop a novel method for determining the class distribution of an unlabeled dataset, along with a fully automated and unsupervised class labeling framework. We validate our methods across a diverse set of real-world tabular datasets that vary widely in domain, class distribution, feature dimensionality, and size, including challenging applications such as fraud detection and cognitive assessment.
Our unique approach involves the combination of two labeling strategies, an unsupervised ensemble and percentile-threshold based methods, that create a high-confidence set of labels which ultimately determine a single positive or negative label for each instance in the dataset based on the expected number of positives. We further improve label quality and efficiency by integrating unsupervised feature selection to rank and identify the most informative features. Unsupervised feature selection simplifies the model and reduces computational complexity, making the method well-suited for large-scale, severely imbalanced datasets (e.g., Medicare and credit card fraud). Moreover, we enhance our labeling method by introducing an unsupervised framework that automatically estimates the class distribution. Using this estimate, the framework selects decision thresholds adaptively, thereby improving label quality. Our novel approach relies exclusively on the dataset’s own features for labeling, requiring no external labels or manual annotations. This makes the method fully automated and unsupervised. We detail empirical results demonstrating substantial improvements in label quality, both across refinements of the method (e.g., progressing from unsupervised to automated unsupervised approaches) and in comparison to an unsupervised baseline learner. These results highlight the effectiveness of our novel class distribution estimation and class label generation methods when applied to unlabeled data
MATERNAL STRESS AND INFANT ANXIETY: EXAMINING THE LINK THROUGH FEAR TEMPERAMENT
Full text linkThis study explored the relationship between maternal stress and infant fear temperament. Infant temperament is important because it allows the caregiver to understand the individual variation in the emotional and social needs of the child. Due to previous research, it can be inferred that maternal stress can influence the development of infant fear temperament. The current study looks at the relationship between maternal stress and infant behavioral fear temperament. In this study there was no relationship between maternal stress and infant behavioral fear when collapsing the across conditions and emotions expressed during the Lab-TAB. However, infants varied in their fear compared to sadness or distress responses across conditions, and this was slightly stronger for the infants of mothers with higher stress levels than those with lower stress vi levels. Overall, the findings demonstrate that maternal stress has a complex impact on infant behavioral responses