41530 research outputs found
Sort by
Frequent Itemset Mining with tidyclust in R
Get PDFUnsupervised learning is closely associated with clustering, however other methods fall under this umbrella such as data mining. In R, the tidyclust package provides a unified interface for clustering models, yet lacks support for data mining. This thesis addresses this gap by introducing the Apriori and ECLAT algorithms into tidyclust, with a focus on frequent itemset mining. Unlike traditional clustering models, frequent itemsets produce groupings of column variables, rather than cluster labels or partitions of observations. To address this, a novel clustering approach is proposed: items (columns) are grouped based on their ”dominant” frequent itemset. A key contribution is a new prediction method, modeled as a recommender system, to predict missing items. This implementation extends tidyclust to support column-based clustering, with applications in market basket analysis and recommender systems
Density-Based and Model-Based Clustering with Tidyclust in R
Get PDFClustering is a fundamental technique in unsupervised learning that can be used to find hidden patterns and structures within unlabeled data. The tidyclust package in R provides a unified interface for applying various clustering techniques to data. This paper outlines the addition of density-based clustering with DBSCAN, and model-based clustering using Gaussian mixture models (GMMs) to the tidyclust package. DBSCAN can be performed using the db_clust() function and makes use of the dbscan package implementation as its engine. GMMs can be fit using the gm_clust() function which makes use of the mclust package implementation. This paper highlights the changes made to these underlying implementations in the process of bringing these methods into tidyclust. This includes changes to the model argument names, how the model is fit on data, and how the model is used to predict on future data
Tread: A Digital Solution for Sustainable Travel
Get PDFThis senior project aims to create an eco-tourism website that promotes sustainable and ethical travel by connecting users with local businesses, communities, and culturally respectful experiences. As mainstream tourism increasingly contributes to environmental degradation and social displacement, this platform will provide responsible alternatives by avoiding over-touristed destinations and favoring underrepresented areas. It will allow users to book flights, lodging, and activities while educating them on the consequences of over-tourism. Through a combination of intuitive user interface, thoughtful user experience, engaging visual design, and curated educational content, the website will serve not only as a booking platform but as a digital advocate for responsible travel. The project will emphasize accessibility, ethical design, and user trust, ultimately encouraging more mindful and sustainable travel behaviors
Effects of Nutrition Education on Metabolic Risk Factors in Male Hispanic Farmworkers
No full textHispanic farmworkers are an underserved and understudied population with disproportionately high risk for type II diabetes mellitus, driven by both genetic susceptibility and compounded effects of lifestyle, occupational, and socioeconomic factors.To our knowledge, this is the first study to assess both body composition and biochemical markers over time in male Hispanic farmworkers receiving a culturally tailored nutrition education intervention. The purpose of this study was to evaluate changes in body composition and blood biomarkers across a six-month hybrid intervention that delivered educational materials remotely while conducting in-person assessments on the farm site.
Body composition and metabolic markers were assessed at baseline, midpoint, and post-intervention. Participants (n=15) received weekly nutrition education and completed in-person assessments that included bioelectrical impedance analysis and fasting blood work.
Results of this study indicated no statistically significant changes in body composition or blood biochemistry. However, there was a 30% increase in fasting insulin and a corresponding increase in HOMA-IR from baseline to post-intervention, suggesting early shifts in insulin sensitivity. Body composition remained relatively stable, with small non-significant increases in BMI and body fat percentage, and a slight reduction in waist circumference.
These findings suggest that although the intervention did not yield significant improvements, it may have helped mitigate worsening health markers during a period of occupational stress and seasonal change. This study demonstrates the feasibility of conducting community-based assessments in this population and supports the need for longer, more interactive interventions to improve metabolic health among male Hispanic farmworkers
The Effect of Food and Temperature Acclimation on the Proteomic Response to Heat Shock in the Mussel Mytilus
No full textThe ecological factors affecting the heat-shock response have been well investigated over the last decades. However, the effect of food availability has not received much attention despite a likely role in modifying stress tolerance limits. Here we acclimated the intertidal mussel Mytilus californianus to four conditions of low–high food, nearshore (low) and aquaculture (high) phytoplankton levels, combined with low-high (20 °C and 30 °C) aerial temperatures. We simulated circadian-entrained circatidal rhythms, before exposing whole mussels to an acute 33 °C 6 h aerial heat stress and allowed for 1 h and 24 h recovery at acclimation conditions. A parallel group in the same experiment was simultaneously exposed to sirtuin inhibitors; although mentioned briefly, this aspect will not the focus in the present thesis.
Our results demonstrate the fundamental role of food availability in attenuating proteomic changes in core carbohydrate and one carbon metabolism, indicating the folate and methionine cycles as central hubs for providing precursors for histone methyltransferases, and producing purines that likely serve as indicators of energy status. We confirm the importance of food availability on oxidative stress responses, especially the scavenging of hydrogen peroxide and the maintenance of reduced glutathione levels. While low food induced the canonical complement of heat-shock proteins, with an emphasis on endoplasmic reticulum chaperones, it was several subunits of the chaperone-containing T-complex, mainly responsible for folding of cytoskeletal proteins, that showed acclimation-specific responses. Similarly, several proteasomal subunits increased across acclimation groups, while the ubiquitin receptor ADRM (Rpn13) varied in response, indicating changes in the type of protease activity. Histone modifiers such as histone kinases and high-mobility groups proteins were prevalent in all groups, specifically in high temperature – low food mussels. Heat stress also induced different splicing factors and heterogeneous nuclear ribonucleoproteins indicating graded differences in alternative splicing among acclimation groups. Within the translational space, changes in ribosomal protein subunits suggest enhanced mRNA selectivity while translation initiation proteins were stable and only elongation factors increased. Cytoskeletal changes indicated actin stress fiber formation in addition to enhanced cross-linking, bundling, and reinforcement of cell adhesion attachments of actin fibers. While we identified members of many signaling pathways, specific increases in p38 MAPK and stress-activated protein kinase (cJun N-terminal kinase) suggest that high aerial temperatures better prepared mussels for an acute heat stress. The MAPK module may also modify prostaglandin synthesis, the role of which we can only hypothesize. While low temperature-low food (LTLF) induced the full complement, both low temperature-high food (LTHF) and high temperature-low food (HTLF) were prepared to limit the response, while high temperature-high food (HTHF) mussels showed an in-between stress response, leaving us to hypothesize there is a trade-off to high food levels in preparation for an acute stress response
Drone Package Delivery System
Get PDFIn response to the demands of the 2025 C-UASC (California Unmanned Aerial Systems Competition), our team developed an autonomous drone payload delivery system capable of two distinct delivery modes: a precision-controlled descent and a high-speed aerial drop. Designed for modular integration with UAV platforms, the system prioritizes reliability, accuracy, and simplicity under desert environmental conditions. Our solution features a custom winch mechanism with a programmable stepper motor, a passive mechanical hook for tension-based release, and a servo-actuated trapdoor for high-altitude drops. The design was extensively tested for structural integrity, motor torque requirements, drop precision, and electronics reliability—culminating in a robust prototype. This project demonstrates how practical engineering design, iterative prototyping, and system-level testing can converge to meet the rigorous demands of real-world challenges