26800 research outputs found
Sort by
Development and usability testing of a preliminary web-based application for the clinical implementation of blood flow restriction: a mixed methods pilot study
No full textIntroduction: Exercise with blood flow restriction (BFR) has gained popularity for use with a wide range of healthy and clinical populations. However, several factors including medical screening, selection of equipment, and determination of cuff pressure still pose barriers for implementation. Accordingly, this study aimed to develop and test a web-based application to guide practitioners in using BFR safely and effectively. Methods: First, we developed an application to assist with medical screening, selection of appropriate equipment, and determination of cuff pressures. Subsequently, we conducted preliminary usability testing of the application using a mixed methods approach. Licensed physical therapists (n = 5) with no prior experience with BFR used the application to implement BFR exercise in hypothetical patient scenarios. Afterward, perceived usability was assessed using the System Usability Scale (SUS) and semi-structured interviews analyzed through thematic analysis. Results: All task scenarios were successfully completed in an average time of 2.3 ± 1.2 min. A total of 11 errors occurred, including minor navigation issues (4), data input problems (2), and difficulty interpreting recommendations (5). The composite SUS score was 94 ± 5, ranking highly compared to industry standards. Interviews revealed that the application was efficient, boosted confidence in using BFR, and increased the perceived likelihood of incorporating BFR into clinical practice. Discussion: These findings suggest that the web-based application has potential to serve as a valuable tool for overcoming barriers to BFR use, enhancing accessibility, and improving the safety and effectiveness of BFR implementation in clinical settings
A regional coastal Douglas-fir index of site quality for young stands in western Washington, USA
No full textSite index is frequently used for deciding how to manage stands, however, it often omits or fails to accurately characterize the site-dependent growth of young (ages ≤8 years) trees. This early stage of development is a critical period in a rotation during which important management decisions with the potential to influence all future growth are made. To address this, we developed a new index of site quality for young, coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii) stands in western Washington, USA. We used high-density light detection and ranging (LiDAR) scans for 161 stands between 3 and 8 years old, in conjunction with Soil Survey Geographic Database information to identify site characteristics that were associated with productivity in young Douglas-fir sites. We utilized individual-tree detection algorithms to measure tree heights and developed a linear model for correcting LiDAR-derived heights in young Douglas-fir stands based on field-measured heights. Our findings indicated that Douglas-fir growth tended to be greatest on sites that were low in elevation, high in precipitation, and had fine-textured soils. We provide an index of site quality for young Douglas-fir based on elevation, precipitation, and soil textural classes to guide early stand management
Examining Video-Based Learning in Higher Education
No full textInstructional video has proliferated recently due to the explosion in popularity of TikTok and YouTube. However, most videos are not developed using principles derived from research and instructional design theories that have been tested with content presented in other types of media (de Konig et al., 2018). Many of these principles are based on the Cognitive Theory of Multimedia Learning, including research on the seductive detail effect; seductive details are “interesting but irrelevant details that are added to a passage to make it more interesting” and have been shown to harm learning (Harp & Mayer, 1997). This has motivated the coherence principle, indicating that learning is more effective when extraneous materials are excluded. Instructional videos by well-known influencers on TikTok and YouTube appear to rely on extraneous content to maintain viewer interest and engagement; it is unclear how effective they are in terms of learning. This dissertation begins with a history of video in education and a theoretical background of the educational concepts around multimedia learning. Study 1 and Study 3 examine learning with instructional videos created by two prominent social media content creators. In both cases, performance was compared to that from videos with matched content but without seductive details; these videos more closely held to best practices supporting the coherence principle. Study 2 surveyed participants’ instructional video preferences and examined the effect of motivation on their preferences and behavior. In studies 1 and 3, material traditionally considered extraneous did not impair learning outcomes and may have contributed to increased engagement. In Study 1, seductive details were noticed and increased extraneous load, but did not hinder recall or transfer. In Study 3, the TikTok-style videos violated the coherence principle, but were rated more favorably and preferred by participants, with no negative impact on recall or transfer. Study 2 revealed that motivation had a clear influence on how learners interacted with instructional videos. These results suggest that the coherence principle should be weighed against learner engagement and motivational factors. Rather than viewing all non-essential content as detrimental, instructional designers may need to distinguish between “distracting” and “motivating” extraneous content, especially in learner-driven or informal learning environments. The dissertation concludes with recommendations for creating instructional videos
Sensing Tap Location Using an Array of Acoustic Sensors
No full textAs virtual and augmented reality (VR/AR) technologies continue to evolve, the demand for more natural and integrated input systems has become increasingly evident. Traditional devices such as keyboards, mice, and touchscreens are often unsuitable for immersive environments due to their reliance on bulky physical hardware and constrained interaction spaces. This work presents an acoustic-based input system that enables ordinary surfaces to function as interactive interfaces by analyzing the propagation of sound generated by finger taps. The system employs sound source localization (SSL) with time difference of arrival (TDoA) algorithms to accurately detect and localize taps on a flat surface. To ensure practical applicability, the system is also designed with a focus on real-time processing and robustness to ambient noise. By addressing key challenges in signal processing and localization accuracy, this work demonstrates a cost-effective and flexible alternative to conventional input methods. The proposed approach has the potential to significantly enhance user interaction in AR/VR environments, offering an adaptable solution for future applications
Using Tunable Diode Laser Absorption Spectroscopy to Measure Atmospheric Water Vapor Concentrations
No full textObtaining accurate measurements of water vapor concentrations is essential for understanding many atmospheric phenomena, such as cloud formation and radiative balance. The ability to do so while minimizing direct contact with the system through remote sensing is especially valuable. With this project, we continued the development of a tunable diode laser hygrometer that provides highly accurate, path-integrated measurements of water vapor concentration. This system utilizes a balanced photodiode circuit to eliminate the power ramp intrinsic to the diode laser wavelength tuning process. The results of this laboratory work show excellent agreement between transmission measurements and simulations based on the HIgh-resolution TRANsmission molecular absorption database (HITRAN). The benchtop system characterized and described here will be packaged into a laboratory instrument to be used with MTU’s Pi Cloud Chamber. Upon completion the hygrometer should be widely applicable to many studies in clouds and aerosol research. The ability to use simulation to bypass much of the need for calibration should also allow for rapid deployment
A DOMAIN-ADAPTED NATURAL LANGUAGE PROCESSING FRAMEWORK FOR MINING SAFETY ANALYTICS: FROM LATENT PATTERN DISCOVERY TO AUTOMATED INFORMATION EXTRACTION
No full textOccupational safety remains a critical global concern, particularly in high-risk sectors such as mining, where analysis of historical accident data is essential for identifying hazards and guiding preventive measures. While structured data has traditionally supported retrospective safety analytics, the rich contextual information embedded in unstructured accident narratives remains largely underutilized in current safety management practice for proactive risk assessment and targeted safety intervention. This dissertation addresses this gap by presenting a comprehensive natural language processing (NLP) framework that addresses fundamental challenges through an integrated approach combining semantic text analysis, domain-adaptive language modeling, and automated safety information extraction for mining safety analytics. Traditional NLP techniques face significant limitations in capturing nuanced semantic relationships within unstructured text and in effectively integrating structured metadata. To address these, the framework introduces a clustering-based semantic analysis framework that combines transformer-based sentence embeddings, nonlinear dimensionality reduction, and k-means clustering. By incorporating structured metadata into the embedding process, this unsupervised approach reveals latent accident patterns not captured by conventional techniques, while highlighting the contextual limitations of general-purpose language models in understanding domain-specific safety language. To address these limitations, the study employs Domain-Adaptive Pretraining (DAPT) of Bidirectional Encoder Representations from Transformers (BERT) and parameter-efficient architecture, A Lite BERT (ALBERT), on a multi-source safety corpus spanning mining, construction, transportation, and chemical processing sectors. The resulting models, SafetyBERT and SafetyALBERT, demonstrate substantial improvements in both intrinsic and extrinsic evaluations, outperforming general-domain and larger models, including Llama 3.1-8B, across multiple safety-specific single-task and multi-task classification tasks. Leveraging these models, the framework is extended to an extractive question answering (QA) system that uses SafetyBERT to automatically extract critical safety information, such as incident causes, work activities, and injury types, from unstructured narratives. To reduce annotation costs, an integrated hybrid active learning (AL) strategy is proposed. The AL cold-start problem is mitigated through a strategic seed selection process using unsupervised, embedding-based clustering, followed by a hybrid querying mechanism that combines uncertainty-based and confidence-based sampling for iterative model refinement, achieving robust performance on the extractive QA task. Altogether, this dissertation contributes a robust NLP framework that advances occupational safety analytics in the mining domain. The integrated approach combining unsupervised pattern discovery, domain-adaptive language modeling, and automated safety-critical information extraction enables proactive hazard management and targeted safety interventions
Faster Range LCP Queries in Linear Space
No full textA range LCP query rlcp(α, β) on a text T[1..n] asks to return the length of the longest common prefix of any two suffixes of T with starting positions in a range [α, β]. In this paper we describe a data structure that uses O(n) space and supports range LCP queries in time O(logε n) for any constant ε \u3e 0. Our result is the fastest currently known linear-space solution for this problem
Choosing between alliances and acquisitions: A knowledge recombination perspective
No full textPrior studies examining the role of knowledge similarity in the choice between alliances and acquisitions have often relied on a narrow conceptualization of a firm’s knowledge base. Moreover, they have rarely investigated the implications of this choice. In this study, we take a more holistic view of a firm’s knowledge base to examine how domain knowledge similarity and architectural knowledge dissimilarity impact the choice between alliances and acquisitions. We further explore how that choice, in turn, changes the firm’s knowledge base. Our findings show that as dissimilarity in architectural knowledge with target firms increases, firms are equally likely to pursue either alliances or acquisitions. However, alliances lead to more significant changes in architectural knowledge than acquisitions. Conversely, firms prefer acquisitions when the target firm has similar domain knowledge, and acquisitions contribute to more substantial changes in the depth and the breadth of domain knowledge. Overall, our results indicate that alliances are more effective for acquiring architectural knowledge, whereas acquisitions are better suited for acquiring domain knowledge. These findings offer managers valuable guidance for enhanced decision-making in knowledge acquisition strategies
The Sharing of Disaster-Related Information on Social Media
No full textThis paper explores how two types of textual features—uncertainty-related features and self-regulation-related features—affect information diffusion amid disasters. We identify four textual expressions (i.e., insight, netspeak, work, and reward) of social media posts that generate negative impacts on the diffusion of information. Against the backdrop of the COVID-19 pandemic, we conducted an econometrical study of COVID-19-related posts collected from Weibo, followed by an experiment, to examine the proposed relationships between the reposting behavior and the four textual expressions. Theoretically, we examined the potential effects of information avoidance on the sharing of information on social media during disasters. This study can improve the predictive performance in future disaster-related studies of social media. Practically, government officers are advised that these features may generate negative impacts on the reposting behavior of those who read their posts and hinder the transfer of official information and policy announcements
Environmental identities of college students reveal potential conflicts and common ground for wildlife conservation
No full text“Conservationist” and “environmentalist” are two prominent environmental social identities often perceived as conflicting, particularly on wedge issues like hunting. While these groups may hold differing philosophies, their beliefs could overlap, revealing opportunities for collaboration in conservation. We examined environmental identities among U.S. college students across 22 states (n = 17,203) from 2018 to 2020, assessing identity, wildlife values, and positions on polarizing issues. Students were classified into four groups: conservationists (8 %), environmentalists (9 %), pluralists (59 %) who expressed both identities, and eco-agnostics (24 %) who expressed neither. Environmentalists, the most diverse group demographically, exhibited mutualistic wildlife value orientations, while conservationists, the least diverse, expressed domination-centered value orientations. Conservationists broadly supported hunting and gun rights, while environmentalists favored animal rights. Despite these differences, both groups scored equally high on conservation caring, and all groups—including eco-agnostics—broadly approved of hunting for altruistic reasons (e.g., ecological benefits, reducing crop damage). Our findings highlight distinct yet overlapping environmental identities shaped by demographic and value-based factors. These identities, while appearing polarized, share relational values (e.g., conservation caring, altruistic motivations) that present opportunities for collaboration. Although based in the U.S., these findings have global relevance and reflect the impact of broader trends (e.g., urbanization) on shifting wildlife values. Understanding environmental identities offers a framework to align conservation efforts across diverse cultural contexts, promoting a more inclusive and unified approach to global conservation challenges