21793 research outputs found
Sort by
Finding Balance: Energy, Wealth, and Health
No full textThis dissertation is composed of three manuscripts (manuscript 1 has been published in Energy Economics, manuscript 2 is currently under review with the Journal of Energy Markets, and manuscript 3 is currently under review with the Obesity Research and Clinical Practice journal) that address problems in energy economics, finance, and epidemiology through statistical modeling.
The first manuscript proposes a covariate-dependent mixture model to describe the behavior of electricity DART spreads (defined as the difference between the day-ahead and real-time prices of electricity). The model incorporates multiple regimes and allows covariates to impact both the frequency and severity of DART spread spikes. Using data from the Long Island zone of the New York Independent System Operator, the model demonstrates a strong fit. Results reveal that including covariates in the severity component of the model is crucial, while mild additional performance is obtained with their inclusion in the frequency component. Neural network-based quantile regression benchmarks are unable to improve performance over our mixture model.
The second manuscript examines the diversification benefits of energy commodities during turbulent periods such as those marked by the COVID-19 pandemic and the Russia-Ukraine war, both of which deeply affected energy markets. Revisiting classical allocation strategies, we incorporate electricity futures—a rarely used asset—alongside crude oil and natural gas futures. Using mean-variance optimization, the diversification benefits are evaluated by combining these energy contracts with the S&P 500. Our empirical approach handles the non-stationarity of returns, volatilities, and correlations. Out-of-sample results show improved performance and diversification, especially during crisis periods.
The third manuscript extends existing dual-energy X-ray absorptiometry-based body composition classifications by introducing additional centile cut-offs to capture tail behavior. Using NHANES (National Health and Nutrition Examination Survey) data, we study the association between these phenotypes and health risks, including metabolic syndrome (MetS), depression, sleep disorders, and comorbidities. Nine phenotypes were identified using quantile regression (QR), and logistic regression was used to assess their relationship with health risks, compared to standard adiposity measures like body mass index (BMI), waist circumference (WC), and total fat percent. The QR model has a better (higher) LR+ (positive likelihood ratio) than the median-split model for MetS and comorbidity but consistently underperforms in LR- (negative likelihood ratio) compared to the median-split model. Both models perform worse than BMI and WC. Whether results differ over time or among certain subpopulations should be investigated
Evaluating the Impact of PCA as a Preprocessing Step in EEG Microstate Analysis
No full textEvaluating the Impact of PCA as a Preprocessing Step in EEG Microstate Analysis Tianhao Ning, MASc
Concordia University, 2025
Electroencephalography microstate analysis is now an important technique for investigating temporal dynamics of brain activity. Given that data are commonly recorded by a large number of electrodes, the high-dimensional nature of this data introduces considerable computational challenges to applying traditional clustering techniques to microstate analysis.
This study examines the application of Principal Component Analysis (PCA) as a dimensionality reduction technique to address the challenge. The PCA projects high-dimensional data into a lower-dimensional space while retaining the most critical features, which can enhance the performance of clustering algorithms like k-means. Firstly, we reproduced the EEG microstate topographies according to Jia & Zeng's work, "EEG signals respond differently to idea generation, idea evolution and evaluation in a loosely controlled creativity experiment," published in 2021, as a baseline. We then compared these results with those of the PCA-reduced data to test whether the structural characteristics of microstates concerning global explained variance, topography similarities, and coverage are preserved after dimensionality reduction.
We tested the performance of PCA in this way, maintaining 95%, 98%, and 99% variance for the original clean data for different cognitive tasks. Our results showed that PCA reduced processing time and maintained the quality of the data, particularly at 98%. However, the PCA-reduced data results differed from those obtained from the original clean data, indicating that PCA is indeed helpful but does alter the results after all.
It also has broad implications for the interpretation of how dimensionality reduction affects EEG microstate analysis, potentially providing superior techniques for handling high-dimensional neural data. Future studies should investigate the biological significance of PCA's effects on EEG data and the impact of PCA on noise signals and test whether this method is suitable as a preprocessing step for other microstate clustering algorithms to ensure that the conclusions reached analytically are both reliable and valid.
Keywords: Electroencephalography, Principal Component Analysis, Microstate Analysis, Clustering, Dimensionality Reductio
Othering in Canadian science textbooks: An analysis of visual, textual and discursive elements
No full textThis study examined 17 science textbooks for grades 7-9 through a mixed-method approach that combined quantitative and qualitative analysis. We analyzed visuals, textual and discursive elements to explore how diversity is represented and whether these materials reinforce or challenge colonial narratives. Grounded in decolonial and post-colonial approaches, feminist science and technology studies and othering, the research applied Content Analysis and Critical Discourse Analysis (CDA) to identify issues of representation and colonial discourses.
Our focus was on how social groups, places and historical times are positioned in the textbooks. Prior studies has emphasized that textbooks function as colonial artifacts, often embedding hidden messages that privilege Western science while marginalizing women, Indigenous peoples, and visible minorities. This research contributes to these debates by demonstrating how textbooks may act as mechanisms of exclusion, shaping perceptions of who produces legitimate knowledge.
Findings show that across images, texts, and discourse, biased representations are reproduced. In images, visible minorities appear statistically overrepresented, yet intersections of gender, race, and role reveal that prominent scientists are almost exclusively White and male. Textual analysis confirms that most scientists highlighted are contemporary, but predominantly European and North American men. Discourse analysis further uncovers colonial logics, including binary oppositions that elevate Western science while erasing or subordinating alternative knowledges.
By exposing these mechanisms of exclusion and authority, this study underscores how science textbooks reproduce social hierarchies. It advocates for efforts to decolonize science education and advance social justice by promoting inclusive and pluralistic understanding of knowledge construction
Determination of factors used to influence purchasing price: Enhancing profitability and competitive edge in discount retail chains
No full textWintana Tadesse – Determination of factors used to influence purchasing price: Enhancing profitability and competitive edge in discount retail chains
This study investigates the key determinants of purchasing prices in the retail sector, with a focus on Dollarama, a major Canadian discount retailer. Using Ordinary Least Squares (OLS) regression, the research analyzes the influence of four core operational variables: Gross Margin (GM), Minimum Order Quantity (MOQ), stock levels, and consumer demand—on purchasing price. Initial exploratory analysis using scatter plots indicated a strong positive relationship between GM and purchasing price, a negative relationship with demand, and weaker but noticeable trends for stock levels and MOQ.
Diagnostic checks revealed violations of OLS assumptions related to linearity and constant variance. To address these issues, a logarithmic transformation was applied to both dependent and independent variables. Post-transformation, the model satisfied all key assumptions, enhancing the robustness and interpretability of the regression results. The refined analysis confirmed that GM, MOQ, and stock levels have a positive and statistically significant effect on purchasing prices, while demand shows a negative effect—suggesting that increased demand may be associated with supplier discounts or economies of scale in procurement.
To capture more nuanced relationships, interaction terms (e.g., stockouts × MOQ, GM × demand) were introduced. These revealed that the effects of some variables are conditional on others, indicating that purchasing price is influenced by interdependent operational dynamics. However, the addition of interaction terms also increased multicollinearity, diminishing the individual significance of previously important predictors. Variance Inflation Factor (VIF) analysis is proposed as a next step to evaluate and address this issue.
Furthermore, the study explores potential endogeneity by regressing current purchasing prices on lagged values of the independent variables. The significance of these lagged variables suggests that past operational conditions have a persistent impact on present pricing decisions.
The findings hold practical implications for retail decision-makers. Understanding how GM targets, MOQ requirements, inventory levels, and demand trends influence purchasing prices enables retailers to develop more informed procurement strategies, negotiate better supplier terms, and optimize inventory management. These insights are especially critical for discount retailers where cost control and pricing efficiency directly impact profitability
Interpersonal Capitalization and Unmet Interpersonal Needs Among Adolescents at Varying Risk for Suicidal Ideation: A Daily Diary Study
No full textAdolescents at risk for suicidal ideation tend to report having unmet interpersonal needs, including perceived burdensomeness and thwarted belongingness. Interpersonal capitalization - an interpersonal process involving disclosing a positive personal event to others and evaluating others’ responsiveness to such disclosure – can promote positive affect and social connectedness. This study examined whether daily capitalization attempts and perceived active-constructive responses to the capitalization attempts were associated with positive affect and fluctuations in unmet interpersonal needs among adolescents at varying risk of suicidal ideation. Adolescents (Mage=15.55; range=12-18) with and without major depression, at respectively higher-risk (n=23) and lower-risk (n=32) for suicidal ideation, completed 10 consecutive daily diaries reporting on interpersonal capitalization, positive affect, perceived burdensomeness, and loneliness (as a proxy for thwarted belongingness). Within-person hierarchical linear modeling analyses showed that daily capitalization attempts and perceived active-constructive responses were associated with higher positive affect in both groups. In the higher-risk group, daily capitalization attempts were associated with lower perceived burdensomeness (b=-0.102, p<0.05) and daily perceived active-constructive responses were related to lower loneliness (b=-0.279, p<0.001). Interpersonal capitalization is a positive interpersonal process that may modulate suicidal risk by shaping perceptions of unmet interpersonal needs in adolescents, particularly among those at higher risk for suicidal ideation
Exploring Efficiency in Split Federated Learning under System Heterogeneity: A Study of System Synchronization and Resource Optimization
No full textThe rise of intelligent edge services has intensified the demand for scalable and privacy-preserving collaborative learning frameworks. Split Federated Learning (SFL), a hybrid paradigm that combines the layer-wise decoupling of Split Learning with the distributed aggregation of Federated Learning, offers an efficient training approach across heterogeneous devices. However, its scalability remains constrained by device heterogeneity and synchronization delays. This thesis proposes a novel Collaborative Split Federated Learning (CSFL) framework that facilitates real-time cooperative computation through direct device-to-device communication. At its core lies the Collaborative Relay Optimization Mechanism (CROM), in which efficient devices act as computational relays by executing intermediate model segments on behalf of bottleneck devices, thereby achieving balanced workload distribution. To minimize end-to-end training latency, we formulate a joint optimization problem over the model cut-layer and device pairing configuration. This problem is decomposed into two interdependent sub-tasks: cut-layer selection, addressed via a convergent alternating optimization strategy, and device pairing, modeled as a deferred-acceptance-based weighted matching game that aligns local utilities with global objectives. Extensive experiments on VGG-16 using the Tiny-ImageNet dataset under non-IID data partitions demonstrate that CSFL significantly reduces training latency and energy consumption while maintaining accuracy comparable to conventional SFL. Overall, this work advances collaborative edge learning by introducing a unified framework for system-level co-optimization, emphasizing the importance of algorithm-architecture codesign for future intelligent systems
Gorge: The Legacy of Grotesque Feminist Performance Art in ‘Park and Scarf’ Mukbang Content on TikTok
No full textThe Western adaptation of mukbang – a Korean media form defined by the excessive
consumption of food on camera – has produced a TikTok sub-genre of content referred to, in this
project, as ‘park and scarf.’ These videos feature young women eating fast or packaged foods in
parked cars and place a sensory emphasis on the process of consumption. This project situates
‘park and scarf’ content within feminist discourses on the female grotesque and Kristevian
abjection, arguing that these performances foreground the unruly body by staging appetite as
both excessive and pleasurable, destabilizing cultural prescriptions of feminine restraint.
The methodology combines feminist performance studies with textual analysis, using a
comparative framework that pairs individual works of performance art with ‘park and scarf’
TikToks. This structure emphasizes how eating, whether staged in an artistic context or
disseminated through social media, mobilizes excess and sensory immediacy and can be used as
a vehicle to interrogate gendered embodiment. By analyzing savoury and sweet consumption
across both artistic and digital performances, the project demonstrates how abjection coupled
with conventional beauty functions as an aesthetic and affective strategy. Excessive female
consumption emerges as a subversive mode that reclaims appetite as a site used to interrogate
and question hegemonic aesthetics of suppression
Occupancy-Informed Energy Management Strategies for Grid-Interactive Buildings
No full textThe global shift towards electrification transforms buildings from passive electricity consumers into active prosumers capable of energy generation, storage, and trading. This evolution expands the building-to-grid (B2G) interactions, moving beyond conventional demand-side management (DSM) strategies toward dynamic peer-to-peer (P2P) energy markets. Such markets necessitate innovative solutions to align energy supply with fluctuating demand through enhanced building flexibility. This thesis focuses on occupancy as a critical but underutilized dimension of building energy flexibility, particularly for informing control strategies in B2G and P2P energy markets. It develops occupancy-informed energy management strategies that dynamically respond to real-time occupancy variations, optimizing energy use and facilitating effective interactions between buildings and the local grid, as well as with peer buildings. The research primary objectives are to: (1) develop approaches for generating representative occupancy schedules for the urban modelling of various building types; (2) formulate occupancy-informed control algorithms for residential and non-residential buildings aiming at enhancing the energy flexibility of these buildings; (3) evaluate the impact of these strategies at both building and grid-distribution levels, addressing roles of both consumers and prosumers; and (4) assess the feasibility of the developed strategies within the current energy policies and future P2P market contexts.
Utilizing open-source datasets, such as mobile positioning data (GPS) and smart thermostat data, combined with advanced modelling, this thesis developed occupancy schedule generators (OSGs) suitable for urban energy simulations. Then, a comprehensive framework was developed to dynamically model urban building energy performance, considering occupancy variations, interactions among buildings, local grids, and neighbouring infrastructure within varying market structures. Applying this framework to urban-scale case studies demonstrated significant reductions in peak energy demand, achieving up to 35% savings in residential buildings and over 17% in non-residential buildings, alongside notable energy cost reductions. Ultimately, this research provides foundational methods and practical tools to leverage building energy flexibility strategically, facilitating optimized interactions within evolving sustainable energy systems
Entanglements of Galois Representations for Elliptic Curves over Q: Foundations and Future Directions
No full textThis thesis investigates the non-surjectivity of adelic Galois representations associated with elliptic curves over Q, a phenomenon explained by two forms of entanglement. We introduce and differentiate between vertical and horizontal entanglements, providing a group-theoretic perspective on the latter. We also develop the concept of 'entanglement networks', diagrams derived from a theorem on field intersections, which offer a framework for analyzing these phenomena and suggest avenues for future combinatorial and cryptographic study. Computationally, we address the classification of potential mod-n Galois images by providing SageMath code to compute all applicable subgroups of GL_2(Z/nZ) for any n. Further SageMath implementations are presented to compute relevant entanglements for a given elliptic curve, utilizing data from the LMFDB database. Finally, we present a theorem providing conditions for the surjectivity of mod-n Galois representations, linking it to the surjectivity of mod-p representations for prime factors p of n and a constant related to Serre's work
Active Grieving: The Aesthetic Activism of ACT UP Montréal
No full textMuch of the historical record on HIV/AIDS activism focuses on the highly documented activities of ACT UP/NY while silencing the work and achievements of other chapters and organizations also operating at the height of the crisis. Following the 1989 AIDS Conference in Montréal, a local chapter christened ACT UP Montréal was co-founded by ACT UP/NY’s Blane Charles the following year. Equally vibrant, the graphic ephemera of ACT UP Montréal—posters, protest signs, pamphlets, manifestation documentation and T-shirts—demonstrates activist voices deserving of similar in-depth analysis. What emerges through looking closely at these items is not only a distinctly québécois framework, but an entirely different set of reference points and goals for HIV/AIDS activist work when compared to ACT UP/NY (often referred to as “the Vatican” in interviews collected for this work). Furthermore, “Active Grieving: The Aesthetic Activism of ACT UP Montréal” reveals a multifaceted microhistory of localized reaction to an international pandemic. This thesis builds on the increasing literature and academic discourse on the many groups that made revolutionary change during this era, this work connects activism history, art history, oral history, archival history, and object analysis to contextualize better known HIV/AIDS activist imagery and show the foundational differences between ACT UP Montréal and other activist groups