South Dakota State University
Public Research Access Institutional Repository and Information ExchangeNot a member yet
30309 research outputs found
Sort by
Generative AI in the Workplace: Adoption Patterns, Innovation Attributes, and Equity Implications at BHSSC
No full textThis case study explores generative artificial intelligence (GenAI) adoption patterns, innovation attributes, and equity implications within a diverse educational cooperative in western South Dakota. A mixed methods approach was employed to analyze adoption trends, perceived productivity differentials, and the significance of key innovation characteristics to 40 employees representing a range of divisions, roles, and levels of digital competence. Employees with lower self-reported digital competence and higher average age reported disproportionately higher perceived productivity gains from GenAI tools. This may suggest GenAI may level the playing field for aging individuals who have previously felt marginalized by rapid technological change in the workplace. Using an extended model of the Innovation Diffusion Theory (IDT) framework, this study also finds support that the attributes of relative advantage, ease of use, result demonstrability, and trialability are significant predictors of GenAI adoption. This has important implications for less digitally confident individuals as well as their employers as it may indicate a path toward unlocking latent productivity potential in the aging workforce
Implications of assuming common within-source distributions and their effect on evidence interpretation
No full textWithin the forensic science community, there is a need for a statistically rigorous determination of whether an exclusionary difference exists; this determination is integral to the Kirkian Two-Stage approach to evidence interpretation (Parker 1966). If the known source is not excluded as the source of the questioned object in the first stage, then the examiner must determine the rate at which competing alternative sources are excluded. Current methods are typically constructed to ensure the same false exclusion rate for each source. For example, in ASTM glass standards E2927-16e1 and E2330-19, an exclusionary difference occurs if any of the standardized differences between the measured trace element concentrations is greater than a fixed threshold of four. However, if the algorithm’s score function has a distribution that varies by source, then the corresponding thresholds will need to vary as well.
In this work, we review strategies for identifying when the within-source distribution of scores varies between sources; methods for estimating thresholds; remedial approaches such as pooling subsets of sources together; and implications of this type of variability among the sources to the Kirkian and likelihood ratio (LR) approaches. We illustrate these methods with example data from traces such as glass and improvised explosive device components.
Although the focus is on the Two-Stage approach, this work is also important for LR-based methods due to the need to estimate a likelihood function from just a few observations from a specified source. The discussed remedial methods also apply to the LR paradigm for evidence interpretation
Evaluating the Effectiveness of OPTN Regulations in Organ Transplantation
No full textOrgan transplantation is often the only therapy available to patients with end-stage organ failure. In 2024, over 48,000 transplants were performed in the US, while another 67,000 candidates joined the national waitlist. The Organ Procurement & Transplantation Network (OPTN) manages the national transplant system and oversees the operation of transplant programs (i.e., hospitals that perform transplants). The OPTN has historically monitored transplant program performance using 1-year post-transplant survival metrics. Recently, the OPTN introduced a new set of criteria with additional evaluation metrics to improve performance monitoring. Our study investigates the effectiveness and possible limitations of these regulations using a simulated environment. As a first step, we aim to quantify the prevalence of ‘false flagging’, i.e., the likelihood that a well-performing transplant program is incorrectly flagged and vice-versa. We develop a Monte Carlo simulation framework comprising three phases: (i) preliminary phase, where we estimate patients’ pre- and post-transplant mortality; (ii) simulation phase, where we assign patients to virtual transplant programs and generate simulated patient outcomes; and (iii) evaluation phase, where we assess the accuracy of the new metrics. Because true program-level risks are unobservable, we artificially designate certain programs as high-risk and simulate patient outcomes using preliminary regression models. The models are calibrated using national data on adult first-time kidney transplant candidates waitlisted between 2012 and 2019. Our findings aim to substantiate the OPTN’s efforts to refine performance metrics, prioritize patient safety, and promote increased utilization of deceased-donor organs.
KEYWORDS: organ transplantation, Monte Carlo simulation, health policy, data science in healthcare
Deep Learning for Forensic Identification of Source
No full textWe used contrastive neural networks to learn a similarity score within the framework of the forensic common-but-unknown source problem. Similarity scores are often used for the interpretation of forensic evidence. Utilizing the NBIDE dataset of 144 spent cartridge casings, this work tested the ability of contrastive neural networks to learn useful similarity scores. The results obtained by contrastive learning were directly compared to a standard forensic statistics algorithm, Congruent Matching Cells (CMC). When trained on the E3 dataset of 2967 spent cartridge casings, contrastive networks outperformed the CMC algorithm. Generally held principles in deep learning would suggest that a larger training dataset would yield even more effective similarity scores. We also considered the effects of varying the neural network architecture; specifically, altering the network\u27s width or depth. This work was in part motivated by the potential to use similarity scores learned via contrastive networks for standard evidence interpretation methods such as likelihood ratios
F.C.W. Kuehn Papers
Get PDFFrank Charles William Kuehn (1884-1970), an influential architect based in Huron, South Dakota, is known for his work on schools, commercial buildings, churches, and residential properties. His designs for public works, including auditoriums, fire stations, and service stations, reflect his impact on the region\u27s architectural development. Notable projects in the collection include the Huron Dairy Products Company, several South Dakota Standard schools, and the First Methodist Episcopal Church in Doland. The plans capture the growth of South Dakota during the early to mid-20th century and highlight Kuehn’s role in shaping the architectural landscape of the region
Dairy & Food Science News, March 2025
Get PDFMessage from the Department HeadFaculty & Staff UpdatesDepartment Faculty Earn AwardsMidwest Regional Dairy Challenge Competition hosted by SDSUSmall-Scale Dairy Processing ProjectRecruitment Updates & OpportunitiesRecent ActivityAlumni Mentorship ProgramSDSU Food Science Research in the NewsStudent UpdatesEvents & Calendarhttps://openprairie.sdstate.edu/dairy_news/1001/thumbnail.jp
Modeling Area Deprivation Index Using Non-Gaussian Fixed Rank Kriging
No full textSocioeconomic disparities shape health outcomes across the U.S., with the Area Deprivation Index (ADI) serving as a key measure of community-level disadvantage. Predicting ADI using Social Determinants of Health (SDOH) like income and health-care access allows for targeted interventions. However, traditional models often ignore spatial patterns, limiting accuracy. This study compares conventional and spatial models to improve ADI prediction. Tract-level ADI and SDOH data were analyzed using six predictors selected via stepwise AIC: median income and distances to five healthcare facility types. Five models were tested: Linear Regression, GAM, GAMLSS, Gaussian FRK, and Poisson FRK. FRK models use low-rank basis functions to efficiently capture complex spatial dependencies in large datasets. Model performance was evaluated using R², RMSE, MAE, AIC, and cross-validation. Linear regression performed worst; GAM and GAMLSS improved results by modeling non-linearity. Gaussian FRK enhanced spatial prediction but oversmoothed local detail. Poisson FRK delivered the best accuracy, capturing both regional and local deprivation patterns. Based on our findings spatial models are essential for analyzing geographically structured data. They capture spatial dependence and distributional complexity, improving prediction and interpretation, unlike traditional methods. Their use supports more accurate location-specific insights in public health and beyond
Integrating Invasion Risk and Habitat Suitability to Guide Invasive Species Management in Wetlands, Lakes, And Rivers
No full textInvasive species disrupt ecosystems by altering habitats, competing with native species, and disrupting food webs. The northern Great Plains is an ecologically important region with many wetlands, lakes, and rivers that are under threat from two invasive species, Bighead (Hypophthalmichthys nobilis) and Silver Carp (H. molitrix; bigheaded carp). These large planktivores change the abundance and composition of phytoplankton and zooplankton communities which alters prey for native mussels, fishes, and, potentially, waterfowl. Bigheaded carps have invaded a portion of the Northern Great Plains, and preventing further spread is a primary management goal. My work assists these efforts by providing tools to improve preventative management actions for bigheaded carp. While prevention is critical, monitoring all waterbodies for early introductions is impractical. My second chapter aims to help managers prioritize surveillance monitoring locations by predicting habitat suitability for bigheaded carp across 53 waterbodies (wetlands, lakes, and rivers) in the region. I used an individual-based model that predicted bigheaded carp survival and growth based on observed environmental conditions. These predictions helped categorize waterbodies from very high to low risk, guiding managers in prioritizing surveillance efforts and ensuring efficient use of resources. Once high-risk locations are identified, managers still need to know where within a waterbody to sample because invasive individuals are often rare and patchy in occurrence. Since individuals commonly seek patches of high-quality habitat, my third chapter used a modeling approach to predict growth rate potential as a fine-scale habitat quality metric. I then assessed the extent to which the location of high-quality habitat changed throughout the year and assessed whether this occurred more in certain habitat types. The location of high-quality habitat patches was consistent through time for wetlands and lakes, whereas these locations were quite variable throughout the year in rivers. These results suggest that it might be beneficial for sampling protocols to be customized for particular habitat types. Overall, these results will assist management efforts to contain bigheaded carp populations by providing data-driven information about habitat suitability that can be used to target management efforts to the most at-risk locations and to efficiently use management resources
SDSU Data Science Symposium, 2025
No full texthttps://openprairie.sdstate.edu/ds_symposium_2025_gallery/1005/thumbnail.jp
SDSU Data Science Symposium, 2025
No full texthttps://openprairie.sdstate.edu/ds_symposium_2025_gallery/1011/thumbnail.jp