TUScholarShare (Temple University)
Not a member yet
10706 research outputs found
Sort by
Modeling Infrequent Events by Integrating Noisy Data From Multiple Sources And Applying Machine Learning to Classify and Predict Disruptions
No full textAnticipating rare yet consequential events, whether sudden line faults rippling across an electric grid or severe storms crippling communities, demands learning systems that function effectively despite missing data, imperfect labels, and extreme class imbalance. In this dissertation, five studies address this challenge by augmenting limited primary signals with supplementary information sources, fusing the heterogeneous evidence within architectures that preserve space-time structure, and validating their effectiveness through careful sensitivity testing.
Two of the papers focus on power-system reliability, where phasor measurement units (PMUs) sample the grid at sub-second rates but are deployed too sparsely to capture every disturbance and are accompanied by incomplete event logs. The remaining three papers target severe-weather prediction, a domain where automatic weather stations often fail during the very conditions they are designed to monitor, and historical archives contain only a handful of the most disruptive events. Across both domains, the work demonstrates that simulation, social or expert text, and numerical forecasts can each serve as a “second lens” through which the learning algorithm detects patterns that would otherwise remain hidden.
In the power-system studies, the central idea is to use physically realistic simulations to supply the ground-truth diversity missing from field data. Real PMU recordings exhibit two intertwined limitations: many fault waveforms are captured far from their origin, making them difficult to distinguish, and the most destructive fault types appear so rarely that they contribute only a small fraction of the training examples. By generating synthetic voltage and current traces for all fault categories on a benchmark grid model, the first paper constructs a balanced, perfectly labeled companion dataset that is then blended with two years of western interconnection PMU observations. This merged corpus feeds conventional machine-learning classifiers that, once trained, recognize multi-phase faults with a level of confidence previously reserved only for the simpler single-phase cases, while overall misclassification rates drop sharply.
The follow-up paper examines how to deploy simulation most efficiently. A systematic sensitivity analysis varies the number, placement, and voltage level of virtual PMUs, revealing that only a strategically selected subset of high-value sensors located near likely fault sites provides most of the predictive benefit. Even when models are trained entirely on synthetic data, they generalize effectively to a distant regional grid with poorly labeled field data, confirming that simulation can substitute for an expensive and error-prone labeling campaign. Together, these two contributions establish a transferable strategy: generate fault scenarios in silico, select placements guided by grid physics, blend them with whatever real data are available, and train a classifier that remains robust even when confronted with unfamiliar topologies or mislabeled events.
Severe-weather prediction involves many of the same challenges, including sensor outages, label sparsity, and skewed data distributions. Three papers explore how human- and model-generated information can help fill the observational gaps. The first weather study treats social media users as impromptu weather observers. Geotagged messages posted during storms are filtered for meteorological language, encoded using a transformer language model, and temporally aligned with automated surface observing system (ASOS) readings. Feeding this combined representation to a bidirectional sequence network predicts warning signatures of oncoming hazards that were largely invisible to sensor data alone, especially in densely populated areas where social signals are strongest.
Recognizing that social media coverage is uneven, the next study focuses on expert storm narratives maintained by national weather agencies. These concise textual descriptions document the timing, location, and impacts of every officially recorded disaster. Converted to embeddings and late-fused with ASOS measurements, these narratives inject human context that remains available even in remote regions with sparse online activity, consistently improving event detection accuracy. The final paper expands the fusion framework by incorporating short-range numerical weather predictions as a third modality. Forecast fields provide a physics-based description of the evolving atmosphere; textual sources contribute situational awareness of ground impacts; and sensors offer real-time verification. A trimodal architecture, reinforced by spatiotemporal feature engineering, integrates these perspectives without requiring perfect alignment, creating a credible multi-hazard early warning tool that maintains strong performance even in harsh, high-latitude environments where conventional methods often fail.
Although the five studies span two application areas, they converge on several common insights. First, supplementary data, whether synthetically generated, crowd-sourced, written by experts, or forecast by numerical models, are indispensable for addressing the blind spots, noise, and imbalance inherent in primary measurements. Second, late-fusion architectures that allow each modality to develop its own latent representation before merging decisions provide the resilience needed when one stream is delayed, missing, or sampled at a different rate. Third, explicitly modeling temporal and spatial relationships through sequence models and physics-informed feature design is essential, as both grid faults and storms are dynamic processes whose signatures unfold over time and propagate across space. Finally, domain-specific sensitivity analyses translate methodological advances into actionable guidance, such as identifying the grid locations where installing an additional PMU yields the greatest diagnostic gain.
This dissertation offers an end-to-end framework for learning under scarcity in safety-critical settings. In electric grids, it delivers a fault detection pipeline that remains effective even when faced with rare multi-phase disturbances or mislabeled event logs. In meteorology, it transforms partially missing observations into robust predictions of storms that previously escaped detection. More broadly, the work demonstrates that when primary data are lacking, the most effective approach is to enlist the supplementary source that most directly fills the gap and allow a spatiotemporally aware fusion model to integrate the multiple modalities.Computer and Information Scienc
The effects of lexical coverage, discourse structure, and text variables on listening comprehension
No full textLexical knowledge, which is an integral part of second language (L2) development, concerns the requisite amount of vocabulary needed to understand a text. The conclusion in the research literature is that the more lexical items learners know, the better they comprehend reading texts and listening passages. Researchers have attempted to establish lexical thresholds or lexical coverage levels for texts, passages, and genres that can be used to inform vocabulary programs for L2 learners. Lexical coverage as a percentage of known lexical items in a text has been researched in both reading and listening fields. However, no researchers have attempted to evaluate different text types in the presence of lexical coverage manipulation. The purpose of this study is to investigate the differential and interactive effects of lexical coverage, text type—monologues and dialogues—and individual and task variables. In this study, lexical coverage was manipulated by replacing lexical items with pseudowords that mimic English phonological, suprasegmental, and grammatical functions. Lexical coverage at 98%, 95%, and 90% coverage was investigated in conjunction with the manipulation of three monologues and three dialogues. The L2 listening texts were presented to Japanese university participants in nine intact classes in a counter-balanced design to control for topic, lexical coverage, and text type. The results of the multiple-choice listening comprehension tests for each task were evaluated using the Many Facets Rasch Model to estimate topic and item difficulty and instrument performance. The participants’ Rasch person ability estimates for their dictation scores and listening vocabulary levels tests were used as covariates in the main analysis investigating listening comprehension performance in the presence of different text types and lexical coverage levels using Generalized Linear Mixed Effect Modelling.
In this study, I examined how varying levels of lexical coverage influence listening comprehension in monologic and dialogic passages. Six authentic monologues and dialogues were adapted to fit the research context, and in the process, 2%, 5%, and 10% of the content words in each passage were replaced with pseudowords. This procedure resulted in 18 audio versions with lexical coverage levels of 98%, 95%, and 90%. After listening to a passage, participants completed 10 multiple-choice comprehension questions and rated their familiarity with the topic. Comprehension scores, topic familiarity ratings, and proficiency were incorporated into generalized linear mixed models to assess the effects of lexical coverage, text type, and individual listening proficiency alongside aural receptive lexical knowledge on listening comprehension.
The most parsimonious generalized mixed effects model, Model 7, contained only significant factors and predictors. The results of this model indicated that the participants were significantly less likely to answer comprehension questions correctly when listening to passages at 90% lexical coverage compared to 95% or 98% coverage. However, there was no significant difference in performance between 95% and 98% coverage. Post-hoc comparisons showed that the impact of lexical coverage was not consistent across all tasks (Dialogue 1, Dialogue 2, Dialogue 3, Monologue 4, Monologue 5, and Monologue 6). Specifically, lexical coverage had no significant effect on comprehension for Dialogue 1, Dialogue 2, and Monologue 6, while it did significantly influence performance for Dialogue 3, Monologue 4, and Monologue 5. While the reasons for the lack of significant effects in some tasks remain unclear, an analysis of Dialogue 2, on which the participants had the lowest comprehension performance, suggested that factors such as the pseudoword replacement protocol, syntactic complexity, and topic theme might have contributed.
The findings also indicated that the text type of the tasks did not significantly affect the participants’ listening comprehension. Although previous research has suggested that dialogues might be easier for listeners due to features such as repetition, confirmation, and negotiation (e.g., Driscoll et al., 2003; Flowerdew & Tauroza, 1995; Fox Tree & Schrock, 1999; Jung, 2003), the literature on task difficulty differences between monologues and dialogues remains inconclusive (e.g., Bavelas et al., 2000; Branigan et al., 2011; Brindley & Slatyer, 2002). This study aligns with other EFL/ESL research, suggesting that task difficulty results from a combination of task, test items, and individual factors, making it difficult to predict or control. A possible explanation for the lack of significance in this study is that the simplification process removed many authentic features of dialogues, such as pauses, repetitions, and discourse markers, which are often absent in monologues.
Additionally, the results showed that topic familiarity was not a significant predictor of comprehension test performance. This finding contrasts with previous studies (e.g., Giordano, 2021; Kostin, 2004; Nissan et al., 1996). One explanation is that the participants might have confused topic familiarity with task difficulty or comprehension, as they rated topics differently on the background questionnaire compared to after listening. Although neither familiarity rating significantly predicted comprehension scores, the participants appeared to interpret the rating scale inconsistently. This suggests that using a single-item, four-point Likert scale to measure familiarity might lack reliability and methodological soundness.
Finally, Model 7 demonstrated that the participants who performed better on the dictation test and aural receptive vocabulary tests were significantly more likely to achieve higher comprehension scores. These findings highlight the utility of the dictation test in distinguishing participants by proficiency when standardized tests are unavailable. Additionally, the Listening Vocabulary Levels Test, which measured aural receptive lexical knowledge, proved to be a valuable predictor given that it explained unique variance beyond the general listening proficiency captured by the dictation scores.
The value of experimental lexical coverage research is discussed considering the corpus and vocabulary research. Suggestions for methodological improvements are also discussed. Finally, practical suggestions are situated within Nation’s (2007) Four Strands framework so that the current study can be of value to ESL/EFL instructors who hope to provide listening materials at varied lexical coverage levels for their students.Teaching & Learnin
Intrinsic factors associated with engagement in healthy lifestyle behaviors for dementia risk reduction in midlife
No full textIntroduction
Intrinsic motivation is critical for dementia prevention but remains poorly understood.
Methods
A total of 347 middle-aged adults completed questionnaires on intrinsic factors for dementia prevention, demographics, dementia risk, and healthy lifestyle behaviors. Latent profile analysis (LPA) grouped participants with similar intrinsic patterns. Subgroup differences in demographics, extrinsic factors, and healthy behaviors were examined.
Results
LPA identified four intrinsic profiles: Profile 1 had low motivation; Profile 2 had high motivation and self-efficacy, but poor dementia knowledge; Profile 3 had moderate motivation; and Profile 4 had low motivation and high apathy. Subsequent analyses showed that profiles further differed on extrinsic factors, demographic characteristics, and engagement in health behaviors. Specifically, Profile 1 had the lowest dementia risk, best sleep quality, and least loneliness; Profile 2 had the highest income, greater dementia risk, highest cognitive activity, and greatest loneliness; Profile 3 had more caregiving experience and moderate engagement in all healthy behaviors; and Profile 4 had lower incomes, the worst health, and lowest engagement in all healthy behaviors.
Discussion
Results identified groups of middle-aged adults with distinct intrinsic patterns who also differed in demographic/extrinsic factors and health behaviors. These profiles may benefit from different types of intervention strategies for dementia prevention.Temple University. College of Liberal ArtsPsychology and NeuroscienceTemple University Libraries Open Access Publishing Fund, 2024-2025 (Philadelphia, Pa.
Dustin Kidd: Pop Culture Freaks [Audio interview]
No full textDustin Kidd published Pop Culture Freaks: Identity, Mass Media, and Society in 2014 with Westview Press. It’s a hybrid work that combines original research, theoretical and methodological perspectives, and some of the features you might find in a textbook, like infographics and recommended readings. Professor Kidd’s focus is on the popular culture generated by the concentrated corporate mass media, whose revenue model is dependent on rising consumption. One particularly striking insight in this interview was Professor Kidd’s analysis of the changes in TV programming over the past 25 years. In his opinion, sitcoms of the 1970s and 1980s – All in the Family, The Jeffersons – prepared young people for their eventual entry into the work world. By contrast, today’s programs prepare young people to fulfill their role as consumers. As the retail sector has become more important to the overall health of the economy, everyone must be encouraged to consume. Corporate popular culture creates the matrix within which individuals are conditioned to continually seek out new products, new adventures, and new identities. Shop until you drop. Dustin Kidd analyzes the role that race, class, gender, sexuality, and disability play in popular culture, not only in the actual cultural artifacts, like TV programs and films, but also in their creation and then consumption. The fracturing of identity, the privileging of some identities over others, and the yearning for wholeness engendered by this phantasmagoria of identity turns us all into freaks. As our sense of inadequacy ebbs and flows, we search out something to buy in order to feel temporarily at peace. But it doesn’t last long. Fred Rowland recorded this interview with Dustin Kidd on November 18, 2014.Temple University. College of Liberal ArtsTemple University. LibrariesSociologyLearning & Research ServicesAudacityAudacit
Urinary Incontinence & Female Athletes Systematic Review Search Strategy
No full textTo identify studies to include or consider for this systematic review, the review team worked with a medical librarian to develop detailed search strategies for each database. The search was developed for PubMed (NLM) translated into Embase (Elsevier), and Web of Science Core Collection (Clarivate Analytics), CINAHL (Ebscohost), SPORTDiscus with Full Text (Ebscohost), Sports Medicine & Education Index (ProQuest), and Cochrane CENTRAL (Wiley),using a combination of keywords and subject headings. A grey literature search included Clinicaltrails.gov(https://classic.clinicaltrials.gov/)and
WHO International Clinical Trials (https://trialsearch.who.int/). The search was not limited. The final search was completed on February 27, 2025.Temple University. College of Public HealthKrausz Library of Podiatric MedicineTo identify studies to include or consider for this systematic review, the review team worked with a medical librarian to develop detailed search strategies for each database. The search was developed for PubMed (NLM) translated into Embase (Elsevier), and Web of Science Core Collection (Clarivate Analytics), CINAHL (Ebscohost), SPORTDiscus with Full Text (Ebscohost), Sports Medicine & Education Index (ProQuest), and Cochrane CENTRAL (Wiley),using a combination of keywords and subject headings. A grey literature search included Clinicaltrails.gov(https://classic.clinicaltrials.gov/)and
WHO International Clinical Trials (https://trialsearch.who.int/). The search was not limited. The final search was completed on February 27, 2025. The full search details are provided in Appendix(___).
A summary of the search results from databases:
PubMed (NLM) from 1809 to 2/26/2025 (48 Results)
Embase (Elsevier) from 1974 to 2/26/2025 (109 Results)
Web of Science Core Collection (Clarivate Analytics) from 1900 to 2/26/2025 (46 Results)
CINAHL (Ebscohost) from 1976 to 2/26/2025 (22 results)
SPORTDiscus with Full Text (Ebscohost) from inception to 2/26/2025 (37 results)
Sports Medicine & Education Index (ProQuest) from inception to 2/26/2025 (14 results)
Cochrane Central (Wiley) from 1996 to 2/26/2025 (18 results)
Clinicaltrials.gov from inception to 2/26/2025 (16 results)
WHO International Clinical Trials Registry Platform (ICTRP) 1900 to 2/26/2025 (13 results)
The search resulted in 323 studies and 94 duplicate studies were found and omitted by the librarian (SB) using the EndNote 20 duplicate identification strategy. This resulted in 229 records to screen from databases or registers. Studies were screened by title and abstract by two blinded and independent reviewers. If a tiebreaker was needed, a third reviewer was called in. This process was repeated for full text article screening and article selection.Rayyan, https://www.rayyan.ai/Rayyan, https://www.rayyan.ai
Faculty Perceptions of Facilitators and Barriers in Holistic Admissions for Graduate Clinical Health Professions Programs
No full textPurpose: To promote holistic admissions, the focus was on graduate clinical health professions faculty’s perceptions of facilitators and barriers in admissions practices. This evaluation sought to add to existing knowledge of how current holistic practices could be best utilized and implemented in the College of Public Health to effectively support staff and students involved in these programs.
Methods: An anonymous online survey was distributed to the graduate clinical programs within the College of Public Health. The invitation to participate in the survey included the link to the survey, which consisted of 17 multiple choice questions. This was aimed at gaining knowledge of faculty’s knowledge, attitude, and practices related to holistic admissions.
Results: Faculty knowledge, attitudes, and practices of holistic admissions were analyzed using a KAP-based survey. Results indicated faculty perceptions and values being positively aligned with CPH’s, but also needs such as increased time, resources, and a more comprehensive model.Temple University. College of Public HealthHealth and Rehabilitation Science
Energy and Development: A Comparative Appraisal of the Impact of Coal Mining in Britain, the United States, and Twentieth-Century Colonial Africa
No full textSince the eighteenth century, mined coal has been an imperative energy source for enhancing agricultural systems, commerce, and industrial development. The mining industry “has become a global activity critical to wealth creation, power structures and…human progress”. While substantial evidence indicates that coal is mined and used in Europe, the Americas, and other regions, including Africa, the impact of coal-fired energy on socioeconomic and industrial development was far more significant in the Northern Hemisphere. This raises a pertinent question: how and why did coal catalyze industrialization in the developed world—Britain and the United States, yet fail to yield similar transformative results in other world regions? This paper engages in a comparative analysis of the impact of mined coal on Britain and the United States, juxtaposing it with the experience of Africa. Using colonial documents, newspapers, and secondary sources, the paper demonstrates that while the use of mined coal in Britain and the United States was multifaceted—used in the iron and steel industry, textile industry, electricity generation, for powering steam engines, and transport technology, the motive for the establishment of collieries in Africa—Nigeria and South Africa was mainly to power railroads and steamships that would facilitate the exportation of the continent’s cash crop and mineral resources to the imperial metropolis in Europe. Given the centrality of energy to development, this paper argues that the intricate interplay of capitalist and colonial legacies, which dictated the contrasting use of coal-fired energy in Britain and the United States compared to Africa, played a crucial role in the divergent trajectories of industrialization between the two regions.Temple University. College of Liberal ArtsHistoryA paper presented at the 30th James A. Barnes Graduate History Conference, which took place March 14-15, 2025 in Philadelphia, PA
Improving Patient Throughput by Implementing an Electronic Interunit Handoff
No full textBackground: Emergency department (ED) boarding—the prolonged holding of patients awaiting inpatient beds—impacts care quality, efficiency, and outcomes. It contributes to overcrowding, higher mortality rates, delayed interventions, and longer hospital stays. Research has linked extended boarding to worsened clinical outcomes, particularly for critically ill patients. Problem: A suburban county hospital has faced increasing delays in patient handoffs due to challenges in reaching observation nurses for verbal reports. Communication barriers have contributed to prolonged wait times between bed assignment and the "ready to move" (RTM) status, resulting in inefficiencies in patient flow and potential impacts on care delivery. Intervention: A standardized electronic handoff system was initiated between a hospital's ED and OBS unit, replacing verbal reports with an Epic Event Log Summary and structured PACC (Pain, Ambulation, Continence, Contact) notes. Using a pre-/post-intervention design, the study measured time from bed assignment to 'ready to move' status, aiming for completion under 30 minutes. Results: Evaluation of the electronic handoff system compared the time from bed assignment to "ready to move" (RTM) status. Statistical analysis, after removing outliers, indicated a significant reduction in RTM times from 51 to 40.97 minutes (p < 0.001). The mean rank decreased from 212.64 before to 166.88 after implementation, demonstrating improved efficiency in patient transitions. Conclusion: Implementing an electronic handoff system improved workflow efficiency and standardized patient reports. Standardized templates and secure chat confirmations improved communication, reduced boarding time, and enhanced patient flow.Temple University. College of Public HealthNursingA poster presented at the 2025 TUHS Nursing Research & EBP Conference, which took place April 24, 2025, in Philadelphia, PA
Assessing the use of mobile alert systems during extreme weather events: A protocol for a systematic and scoping review
No full textIntroduction/Background: Natural disasters are increasing in frequency and intensity causing damage to the environment, infrastructure, and increased loss of life. Disaster risk communication has a critical role to play in mitigating the impact and damage from instances of severe climate or weather conditions such as heat waves, tropical cyclones, and floods. Short message services (SMS) are often used to issue alerts and warnings ahead of and during an event, providing key information to the public about protective measures.
Objective: The aim of this systematic and scoping review is to assess the literature and understand the use of mobile alerts or warnings used in communicating about extreme weather events. The purpose is to provide emergency preparedness, public health, and communications professionals with an overview of the current landscape to implement this strategy.
Methods: We will identify peer-reviewed journal articles from several online databases including PubMed, Web of Science Core Collection, Academic Search Complete, and GreenFILE. We will first complete a title and abstract search, followed by a full-text screening of the selected articles. Studies will be included if they address natural disasters, e.g. hurricanes, floods, wildfires; extreme weather events, e.g. extreme cold weather event, extreme heat event; the use of a mobile alert or warning strategy, e.g. SMS alerts. Planned search dates are April 2025 through May 2025.
Ethics and Dissemination: This systematic and scoping review focuses on peer-reviewed articles and does not require ethical review. Findings will be disseminated in publications in peer-reviewed journals and conferences.Temple University. College of Public HealthTemple University (Health Sciences Center Campus). LibrarySocial and Behavioral SciencesA poster presented at the Temple University College of Public Health Research and Evidence Based Practice Day, which took place in Philadelphia, PA
Addressing organizational barriers to participation in after-school programming: A collaborative capstone project in partnership with the YWCA Bucks County
No full textThis capstone project was completed working in collaboration with the YWCA of Bucks County and worked to address organizational barriers to participation in after-school programming. Based on initial staff interviews, the project that initially appeared to be most effective fostered more effective communication between YWCA staff and parents. However, following a needs assessment conducted during Phase One of the project, the barriers identified during the needs assessment differed significantly from those identified during the preliminary interviews. Based on these findings, the focus of the project shifted. These barriers were categorized into Context and Performance Patterns, based on the Occupational Therapy Practice Framework (OTPF) and interventions were developed to address both of these areas. Interventions focused on multiple interconnected environmental barriers, and broader organizational barriers to communication, ultimately developing a set of recommendations for YWCA staff to guide future steps.Temple University. College of Public HealthHealth and Rehabilitation Science