Open Research Oklahoma (Oklahoma State Univ.)
Not a member yet
42035 research outputs found
Sort by
User defined interpretability of machine learning algorithms applied to financial time series data
No full textUser studies have be underrepresented in Explainable AI (XAI) literature. While many papers seek to understand interpretability through Folk Psychology or Theory of Mind explanations, the XAI literature also needs to continue to explore rigorous measures of interpretability from the user standpoint. To accomplish this, we employed a straightforward, accurate learning model for automated allocation of capital from financial time series data, and investigated user reactions to, and ability to interpret, varied explanation strategies. After an initial survey gauging a user’s experience with machine learning models in general, the user engaged in an interactive demonstration with the model. Then, we measured the user’s notions of their own understanding and the model’s interpretability and trustworthiness. Interestingly, text explanations were the only type of explanation that increased user trust in a statistically significant way. At the same time, historical examples pulled from the dataset actually negatively impacted user understanding, and were the only explanation type to do so
Effects of maternal influence and season on immune phenotype and stress response of calves from selected sires
No full textIn recent decades, intensified selection pressure on livestock for economically important traits, such as growth, has led to a decline in overall robustness. Beef cattle play a vital role in the U.S. food system and economy; however, they frequently encounter various stressors stemming from environmental conditions, including heat stress, nutritional deficiencies, pests, pathogens, and management practices like handling, weaning, transportation, and commingling. These stressors can adversely affect animal health, well-being, and productivity. Furthermore, individual animals perceive stress differently due to differences in genetic makeup and past experiences, which can lead to varied effects on other biological functions, including the immune system and gut microbiome. It is essential to gain a deeper understanding of how environmental and management-related stressors affect the immune phenotype of beef cattle. Such insights could guide breeding strategies considering paternal production traits and other factors affecting the offspring’s ability to thrive across diverse environmental conditions. The first study evaluated the stress responses and immune phenotypes in steer calves sired by bulls selected for superior growth and milk traits, considering the calving season, from initial processing through a month in the feedlot. The second study further investigated stress responsiveness by utilizing additional stress biomarkers alongside cortisol in steer and heifer calves. The third study focused on a subset of steer calves to analyze gut microbiota using 16S rRNA sequencing of fecal DNA, examining the effects of genetic selection, maternal factors, and calving season post-weaning through the 60-day preconditioning phase. Finally, building on these insights, the last study investigated the interplay between maternal characteristics and sire growth traits during both spring and fall calving seasons to assess their collective impact on immune phenotype, stress responsiveness, and maternal role in vaccine efficacy. These findings revealed that rigorous selection for production traits, maternal factors such as previous experiences, and seasonal challenges differentially impact calf stress responses, immune phenotypes and gut microbiomes. Although further research is necessary to elucidate the underlying mechanisms and other environmental conditions, future breeding strategies should integrate additional biological elements, including maternal traits, production goals, and seasonal factors, to better support calf health and well-being
Embracing culture in STEM: A qualitative examination of a culturally relevant informal learning experience
No full textMathematics and STEM are often viewed as culture-free by students and teachers alike. These types of views can help to alienate students, especially those from underrepresented backgrounds. However, drawing on students’ funds of knowledge by implementing theories, such as culturally relevant pedagogy (CRP), has shown to have positive affective benefits for students. Research has shown mixed results regarding the mathematics- and STEM-related affective benefits of implementing CRP. However, these studies differ in demographics and methods from the current one. Students often form long-lasting attitudes towards mathematics and STEM while they are in the middle-grades. Therefore, the purpose of this study is to qualitatively examine middle-grades students’ interest in and anxiety towards mathematics and STEM. The results demonstrate mixed results regarding how effective program activities were at triggering students’ interest. Additionally, the data indicates the potential for shifting students’ mathematics and STEM-related affective measures. Moreover, the results demonstrate that students’ learning during the program aligned with the tenets of CRP. Therefore, these findings suggest that students’ affective measures can be affected in a short-term informal learning experience. Additionally, the findings suggest that students do not need to participate in long-term programs to learn in a manner consistent with the goals of CRP
Orbital body embeddings: Solitary, intercept, collision, and post-impact debris cloud behavior
No full textUnderstanding and modeling orbital motion is critical for engineering applications. Orbital motion in the two-body problem while well-defined is critical in desiring solutions for orbital debris removal. This thesis focuses on the motion of particles in atmospheric and gravitational orbits and works to provide an understanding of the motion in these fields. The work here focuses on orbital mechanics in low-Earth orbit to provide an understanding of the current orbital debris environment. Estimating the conjunctions of trackable debris for satellites in low Earth orbit. Specifically focused on orbital companion satellites to protect a primary satellite in comparison with statistical engineering models such as ORDEM. In addition to defining the relative motion between two satellites in orbit, for estimation, and non-cooperative rendezvous control, which is essential for debris removal. Finally, the orbital mechanics and relationships are defined to generate theoretical debris clouds from hypervelocity impacts (HVI) in LEO. In addition to providing the tools to estimate where these collisions may occur for any relative velocity, (barv_r). Further, we provide a framework that rotates the HVI debris cloud motion to the Hill frame for the identified collision. Allowing for the rotated HVI data to be the initial conditions for a simulation using the relative equations of motion between two satellites. Thereby, simulating the debris cloud propagation using orbital mechanics and determining the change in orbit, and the resulting risk to satellites in the region human habitable region. This study specifically investigates the effect of debris cloud generation for ISS-like orbits from 400-500km and found in most cases 89% of the resulting cloud de-orbit within 1000s. Integrating the numerical ANSYS tool and classical orbital mechanics, with the relative motion, provides a setup to model short-term orbit cloud propagation and de-orbit simulation
Bacterial small noncoding RNAs affecting gene expression in eukaryotes
No full textWolbachia is a highly studied bacteria found in over 70% of insects around the world. Based on recent research in Dr. Hagen’s lab he suggests that Wolbachia bacteria causes disruption in host’s reproduction abilities due to small noncoding RNAs binding to the host's DNA. This phenomenon is known as cytoplasmic incompatibility (CI) and theoretically affects the way the host expresses genes. Similar research conducted by a fellow undergraduate student suggests the small RNAs produced by Wolbachia bind to Drosophila protein coding genes. This will cause a derangement in the protein expression of the Drosophila helping to explain cytoplasmic incompatibility. This project has allowed for further investigation into the relationships between bacterial small noncoding RNAs with protein coding genes in insects. Previous research in the Hagen Lab has shown evidence that several small noncoding RNAs could have binding potential to protein coding genes. The RNA library pools from Drosophila cells and Drosophila cells infected with Wolbachia have been sequenced and analyzed. The analysis shows several potential candidate 3’UTR sequences. The plasmid used for this project is the pAc5-DsRed2-Sv40 plasmid, which was constructed by a fellow member of the lab from using both the pHd-DsRed plasmid and the act5c plasmid. The pAc5 plasmid contains an ACT5C promoter, the DsRed fluorescent gene, and an ampicillin resistance gene. Different 3’UTR sequences have been added to the plasmid to compare hybridization rates of the small noncoding RNAs. Each plasmid type will be transfected into Drosophila cells, one set infected and the other uninfected. If the small noncoding RNAs of the Wolbachia hybridize to the Drosophila protein coding genes then it is predicted to see an absence or decrease in the fluorescent protein expression. We will be able to recognize the small RNA as nontargeting if the small RNAs are not hybridizing or are low in numbers. With the use of different 3’UTR sequences we will be able to strengthen hybridization of different small RNAs to the Drosophila protein coding genes, reducing the amount of non targeting RNAs. This project aims to uncover the genetic interactions between bacteria and their eukaryotic hosts, helping us understand how these relationships impact human and animal healt
Design and development of an axisymmetric variable area nozzle for small turbojet engines
No full textThis paper presents the design, manufacturing, and testing of a variable area nozzle for small turbojet engines, specifically defined in this study as engines with a mass flow rate lower than five pound-mass per second. The goal is to optimize thrust performance and enable rapid thrust modulation without engine spooling delays, addressing a critical need for precise thrust control in unmanned aerial systems and other propulsion applications. The JetCat P100-RX turbojet engine was selected as the baseline for analysis, with performance targets including a thrust spoilage range exceeding 8-lbf at 80% throttle and a maximum thrust exceeding 19-lbf at full throttle, all while minimizing weight and size constraints. To characterize the performance of the engine, an On-Design Parametric Cycle Analysis and Off-Design Engine Performance Analysis were conducted to determine the optimal and spoiled nozzle exit areas using the Mass Flow Parameter. Several nozzle actuation concepts, including feathered, iris valve, and slotted square designs, were evaluated based on manufacturability, precision, and thermal expansion considerations. Baseline testing the JetCat P100 engines confirmed theoretical thrust predictions and guided the selection of high-temperature-resistant materials. The down-selected concept, a circular rack-and-pinion feather mechanism, was selected for further refinement. Finite Element Analysis and Computational Fluid Dynamics were performed to validate the structural integrity, thermal performance, and aerodynamic characteristics of the design. An iterative prototyping approach was implemented, incorporating outsourced components such as ball linkages, servo motors, and metal-printed nozzle feathers to ensure robust actuation. The final design achieves the required thrust modulation range while maintaining structural reliability and minimal weight. This study demonstrates the feasibility of a variable area nozzle for small turbojet engines, offering insights into future propulsion applications and potential commercialization. By enhancing thrust modulation capabilities, the system provides improved performance for unmanned aerial system applications and military aviation programs. Future work includes physical testing under operational conditions to validate computational predictions and further optimize nozzle responsiveness
Effects of generative artificial intelligence on creative processes within local advertising agencies.
No full textThis qualitative study investigates the perceptions and experiences of advertising and public relations (PR) practitioners concerning the integration of Generative AI (GenAI) tools and their influence on creative workflows in local agencies. With rapid advancements in AI technologies, the advertising and PR industries are increasingly adapting to new creative paradigms shaped by automation, data-driven insights, and the strategic capabilities of GenAI. To understand these shifts, seven industry professionals participated in semi-structured interviews, sharing rich, firsthand accounts of how GenAI supports, challenges, and transforms their day-to-day practices. Through a thematic analysis, five major findings emerged: (1) GenAI enhances human creativity by serving as a brainstorming partner and ideation accelerator; (2) efficiency complements creativity, as task automation allows professionals to focus on higher-order thinking and conceptual development; (3) ethical concerns and data privacy issues influence how and when GenAI is used; (4) training and customization play a crucial role in optimizing GenAI outputs; and (5) GenAI is increasingly viewed as a strategic tool for competitive advantage, predictive analysis and collaboration in a competitive landscape. Overall, the findings underscore that while GenAI is a powerful and evolving tool within the creative domain, its value lies in augmenting—not replacing—human ingenuity. Practitioners view it as a collaborative partner that must be managed with intentionality, transparency, and ongoing learning. The study contributes to broader conversations around the future of work in creative industries, offering both theoretical insights into human-AI collaboration and practical considerations for agencies navigating this technological shift
Injury risk assessments in collegiate female athletes
No full textOnly six percent of sport and exercise science research focuses exclusively on women. Further, factors such as the menstrual cycle, hormonal contraceptive (HC) use, and disordered eating behavior may influence sex-specific musculoskeletal injury, as seen in conditions of the Female Athlete Triad. PURPOSE: The aims of this dissertation were to 1) assess the relationships between menstrual status, hormonal contraceptive use, and skeletal and body composition measures in elite cross-country and track and field (XC&T+F) athletes, 2) examine the prevalence of symptoms relating to the Female Athlete Triad across elite athletes of various sports: XC&T+F, cheerleading, pom, and softball, and assess the most appropriate predictors of injury within each sport, and 3) investigate the relationship between menstrual status and skeletal measures in elite softball athletes, emphasizing practical monitoring protocols and injury predictors to enhance athlete health. METHODS: Elite level athletes completed training, injury and menstrual history questionnaires, the EDE-Q 6.0, validated sleep surveys, a dietary recall, and Dual-energy X-ray Absorptiometry scans (DXA). Statistical analyses were conducted specific to each research project’s aim. RESULTS: This dissertation showcased XC&T+F athletes with infrequent/absent menstrual cycles had lower lumbar spine bone density than athletes using HCs, all teams were at a similar high risk for symptoms relating to the Female Athlete Triad, and the prevalence and impact of the symptoms of the Female Athlete Triad appear to expand beyond only endurance or aesthetic-based sports as softball athletes had a high-risk of symptoms. However, these power-based athletes are not typically considered in this body of research. CONCLUSION: Taken altogether, these findings suggest a more comprehensive athlete health monitoring protocol may be warranted for early diagnosis of Female Athlete Triad symptoms
Conversion in Cusco: The Virgin of Almudena
No full textA statue of the Virgin Mary housed in Cusco weaves an intriguing web of social narratives. She is a copy of the Virgin of Almudena, a miraculous statue which was recovered during the 1085 conquest of Madrid by King Alfonso VI. Found intact behind a wall with two candles still lit despite being hidden there for centuries, she was undisturbed throughout Islamic rule over Spain. By commissioning a copy of this statue in Cusco and placing a splinter of the original inside, not only was Bishop Manuel de Mollinedo y Angulo activating this new statue as a reliquary but was also asserting his status as over his new diocese as an influential and powerful member of the Spanish clergy. This also tethered an older Christian legacy to Andean worship as part of conversion efforts in the Spanish Americas. Executing the commission was indigenous artist Juan Tomás Tuyru Túpac, who completed many religious sculptural commissions in Cusco.
This thesis analyzes the Virgins’ narrative web to trace continual developments of the visual language of the divine in colonial Cusco and the unwieldy concept of copies in Spanish America. The Virgin of Almudena is often mentioned in both Spanish and English scholarship but rarely deeply analyzed. Throughout investigating the purpose of reliquaries and their similarity to Andean huacas and the creation of Marian statues by indigenous artists in their regional devotions this statue serves as a valuable case study of the legacy of colonial tactics of conversion in the Andes and how Christian traditions were emulated and transformed by those being converted
Design and development of a Smart Remote Energy Assessment Framework: A technology-driven approach
No full textRemote energy assessments have emerged as a promising solution to overcome the limitations of conventional on-site energy audits, which are often costly, time-intensive, and logistically challenging. This thesis, titled “Design and Development of a Smart Remote Energy Assessment Framework: A Technology-Driven Approach," suggests the creation of an all-inclusive framework that would increase the effectiveness of auditing processes for industrial and commercial buildings while also maintaining a very high level of accuracy and cost-effectiveness.
This study, guided by an extensive literature review and an examination of current market trends, highlights considerable limitations of conventional assessment approaches, including high operational costs, limited resources, and scalability issues. To overcome these limitations, a mixed-methods research design was used, blending qualitative input from industry experts and quantitative input from market surveys. The main elements of the resulting framework include the development of both a Skill Matrix and a Risk Matrix. The Skill Matrix classifies energy assessment tasks based on the level of expertise required, differentiating between activities that can be competently performed remotely by facility staff and those that require on-site attendance. Concurrently, the Risk Matrix evaluates the reliability of various data collection methods, ensuring that alternative remote data sources yield acceptable accuracy in energy savings estimates.
Validation of the framework was conducted through simulated audits using historical data from Oklahoma State University’s Industrial Training and Assessment Center (OSU-ITAC), and the results show that remote audits are as effective as traditional audits in delivering estimates and cost savings. In addition, it was observed that audit times are decreased by around 45–50% with deviation in savings estimates within acceptable limits of ±5–10%. The results highlight the potential for the use of a hybrid model that incorporates remote assessments with selective on-site inspections for activities demanding higher levels of technical sophistication.
The study concludes that the proposed Smart Remote Energy Assessment Framework offers a viable, scalable, and cost-effective alternative to traditional energy audits and can be used as standardized framework for energy assessment process. Future research directions include the enforcement of real-time auditing using remote means, the enhanced use of Internet of Things sensors and digital twin technologies, and the application of artificial intelligence in order to enhance analysis and decision-making processes. Overall, the study makes strong foundations for the digitalization of the auditing process for energy, hence making the process sustainable and more efficient in the built environment