Theses, Dissertations, and other Required Graduate Degree Essays

Receive updates for this collection

This collection contains digitized SFU theses except for those theses submitted within the last 12 months. If you cannot find the thesis you are looking for please search Recently Submitted Theses as it may be a recently submitted thesis and thus not yet available in Summit.

Approximate marginal likelihoods for shrinkage parameter estimation in penalized logistic regression analysis of case-control data

Author: 
Date created: 
2020-04-17
Abstract: 

Inference of associations between disease status and rare exposures is complicated by the finite-sample bias of the maximum likelihood estimator for logistic regression. Penalised likelihood methods are useful for reducing such bias. In this project, we studied penalisation by a family of log-F priors indexed by a shrinkage parameter m. We propose a method for estimating m based on an approximate marginal likelihood obtained by Laplace approximation. Derivatives of the approximate marginal likelihood for m are challenging to compute, and so we explore several derivative-free optimization approaches to obtaining the maximum marginal likelihood estimate. We conduct a simulation study to evaluate the performance of our method under a variety of data-generating scenarios, and applied the method to real data from a genetic association study of Alzheimer's disease.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Brad McNeney
Department: 
Science: Department of Statistics and Actuarial Science
Thesis type: 
(Project) M.Sc.

Human gait monitoring using wearable fabric-based strain sensors and deep supervised learning

Author: 
Date created: 
2019-12-05
Abstract: 

Continuous lower body monitoring is an important step for real-time feedback training of runners and in-home rehabilitation assessment. Optical motion capture systems are the gold standards for gait analysis, but they are spatially limited to laboratories. Recently, wearable sensors have gained attention as unobtrusive methods to analyze gait metrics and health conditions. In this study, a wearable system capable of estimating lower body joint angles in sagittal, frontal, and transverse planes during gait was developed. A prototype with fiber strain sensors was fabricated. The positions of the sensors on the pelvis were optimized using a genetic algorithm. A cohort of ten people completed 15 minutes of running at 5 different speeds for gait analysis by our prototype device. The joint angles were estimated by a deep convolutional neural network in inter- and intra-participant scenarios. In intra-participant tests, root mean squared error (RMSE) and normalized root mean squared error (NRMSE) of less than 2.2° and 5.3 %, respectively, were obtained for hip, knee, and ankle joints in sagittal, frontal, and transverse planes. The RMSE and NRMSE in inter-participant tests were less than 6.4° and 10%, respectively, in the sagittal plane. The accuracy of this device and methodology could yield potential applications as a soft wearable device for gait monitoring.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Carlo Menon
Department: 
Applied Sciences: School of Mechatronic Systems Engineering
Thesis type: 
(Thesis) M.A.Sc.

Learning shape-to-shape transformation

Author: 
Date created: 
2020-05-14
Abstract: 

Many problems in computer graphics and geometric modeling, e.g., skeletonization, surface completion, and shape style transfer, can be posed as a problem of shape-to-shape transformation. In this thesis, we are interested in learning general-purpose shape transform, e.g., between 3D objects and their skeletons, between chairs and tables, and between letters of two different font styles, etc. With a point-based shape representation, we explore the problem of learning general-purpose shape-to-shape transformation, under two different settings: i). having shape-level supervision, ii). unsupervised. We present P2P-NET, a deep neural network, for learning shape transform under shape-level supervision. It is trained on paired shapes from the source and target domains, but without relying on point-to-point correspondences between the source and target point sets(i.e., point-level supervision). The architecture of the P2P-NET is that of a bi-directional point displacement network, which transforms a source point set to a prediction of the target point set with the same cardinality, and vice versa, by applying point-wise displacement vectors learned from data. For an unsupervised setting, we introduce LOGAN, a deep neural network aimed at learning general-purpose shape transforms from unpaired shape domains. It consists of an autoencoder to encode shapes from the two input domains into a common latent space, where the latent codes are overcomplete representations for shapes. The translator is based on a generative adversarial network (GAN), operating in the latent space, where an adversarial loss enforces cross-domain translation while a feature preservation loss ensures that the right shape features are preserved for a natural shape transform. We conduct ablation studies to validate each of our key designs and demonstrate superior capabilities in shape transforms on a variety of examples over baselines and state-of-the-art approaches. Several different applications enabled by our general-purpose shape transform solutions are presented to highlight the effectiveness, versatility, and potential of our networks in solving a variety of shape-to-shape transformation problems.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Hao (Richard) Zhang
Department: 
Applied Sciences: School of Computing Science
Thesis type: 
(Thesis) Ph.D.

Decolonizing municipal heritage programs: A case study of the city of Victoria’s heritage program

Author: 
Date created: 
2020-04-15
Abstract: 

This research examines the City of Victoria’s heritage program, which comprises of civic plans, policies and associated agencies, to understand whether or not it can meet the Truth and Reconciliation Commission of Canada: Calls to Action. I have argued that the City of Victoria’s heritage program must be adapted to include intangible cultural heritage to support decolonization and the representation of Indigenous cultural heritage. Intangible cultural heritage offers an accessible way for the field of municipal heritage planning to become more inclusive and supportive to reconciliation.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Meg Holden
Department: 
Arts & Social Sciences: Urban Studies Program
Thesis type: 
(Project) M.Urb.

Access, equity, and ethics: A qualitative exploration of Rwanda’s maternal community health worker program

Date created: 
2020-04-28
Abstract: 

Improving maternal health outcomes is one of the main health concerns in Rwanda, a country that was shaken by the 1994 genocide against the Tutsi. As part of the rebuilding process, the health sector focused on using community participation to promote access to maternal healthcare. One such initiative was the creation of the maternal community health worker role as part of the community health worker program. Maternal community health workers are volunteer women elected by their communities to provide basic maternal health services while encouraging the utilization of formal healthcare services for antenatal care, delivery, and postpartum care. Using a qualitative case study approach, my dissertation research explores some of the facilitators and barriers to access to the community-based services offered by maternal health community health workers. I draw on the findings from in-depth interviews with maternal community health workers and women who have used their services in five Rwandan districts to pursue three distinct, yet related, analyses. First, I highlight the different aspects of access to maternal health care at the community level in Rwanda: availability, accessibility, affordability, acceptability, and accommodation. Second, I identify specific strategies employed by these volunteer health workers to facilitate equitable access to maternal health services while operating in a low resource setting. Third, through the lens of an ethics of care framework, I examine why women decide to become maternal community health workers and how they are selected in their communities to take on this responsibility. Overall, this research suggests that community participation is valuable for promoting maternal health outcomes but raises health equity concerns for the nature of the maternal community health worker role. Such concerns shape the program’s sustainability and may impact the overall efforts to enhance positive maternal health outcomes in Rwanda. Further research is needed to explore other aspects of community participation in maternal health, such as the involvement of local leaders who work closely with maternal community health workers to enhance the success of this program.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Valorie Crooks
Department: 
Environment: Department of Geography
Thesis type: 
(Thesis) Ph.D.

Building labour force resilience in British Columbia

Author: 
Date created: 
2020-04-16
Abstract: 

With the rapid growth of automation and technological advancement, the skills and competencies required across British Columbia’s economic development regions are evolving. As the province shifts towards a more digital, knowledge-based economy, it is important to consider the development of BC’s labour force. While there are a number of initiatives targeting the next generation of workers, few supports sufficiently address the needs of mid-career workers in medium-skill occupations, who are more likely to experience challenges in adapting to changing job requirements. The purpose of this study is to determine the role the provincial government can play in building labour market resilience among this group. Using a case-study analysis as the primary research methodology, this study evaluates public employment supports in Ontario, Québec and Australia to identify policy options that may aid in streamlining job-transitions in BC.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Nancy Olewiler
Department: 
Arts & Social Sciences: School of Public Policy
Thesis type: 
(Project) M.P.P.

Experiences of Latin Americans seeking professional jobs in Greater Vancouver

Author: 
Date created: 
2020-03-05
Abstract: 

Canada is often acknowledged as one of the most welcoming countries for immigrants around the world. However, literature reveals that Canadian skilled immigrants, particularly those from Latin America, are often unemployed, underemployed, and earn significantly less than their Canadian-born counterparts. This dissertation examines the experiences of Latin American Skilled Immigrants (LSIs) in Metro Vancouver, including: the factors that prompt them to migrate; their experiences with the Canadian immigration system; and their transition into the new social space. I critically deconstruct dominant economic approaches to immigration and challenge human capital explanations of the phenomena. By utilizing a multiple case study research design, I conducted in-depth interviews with nine LSIs and coauthored their narratives. Filtered through the lenses of Bourdieu's theory of social reproduction, Rizvi’s ideas regarding the neoliberal imaginary, and Bauman’s concepts of the stranger’s aporia, I found that migration appears as a strategy of social reproduction in which participants aim to maintain or enhance their position in the social space. Furthermore, the neoliberal imaginary in conjunction with the participants’ habitus largely shaped their perception of what moving in the social space looks like and how it is achieved. With respect to their transition into Canada, I found that participants who entered with prearranged jobs (WPJ) had more positive experiences settling and adapting than those who entered without prearranged jobs (WOPJ). Participants WOPJ faced more onerous immigration processes and upon arrival, they encountered a contradictory society that intensely seeks to select the best and brightest, but does little to facilitate their integration and in some cases is even obstructive and discriminatory. Through the same theoretical framework, I realized that settling into the community and transitioning into the labour market did not solely depend on the participants’ intrinsic human capital, but also on a complex series of internal contradictions and relations of power created by the neoliberal imaginary. Acknowledging this complexity may lead to a more comprehensive and unprejudiced construction of the Canadian immigration system. This would allow more room to discuss and address the ethical and moral challenges that many immigration stakeholders face, particularly the higher education system in the era of academic neoliberalism.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Kumari Beck
Department: 
Education: Faculty of Education
Thesis type: 
(Thesis) Ed.D.

Survivorship and life history strategies in relation to migration distance in western and semipalmated sandpipers in Perú

Date created: 
2020-03-31
Abstract: 

This thesis explored the relationships between life history, migration distance, survivorship components of fitness, and molt strategies of Western and Semipalmated Sandpipers in one of the most austral non-breeding sites for both species, at Paracas, Perú. I asked how migration distance relates to pre-migratory preparation, survivorship and migratory decisions for different age classes and ecological circumstances between species and within populations. I focused particularly on how timing of first breeding relates to survivorship and thus future overall fitness. I found that adults from both species prepare for northward migration, but no juvenile Western Sandpipers did so, confirming a non-migratory over-summering ‘slow’ life history strategy for more southerly non-breeding populations. Juvenile Semipalmated Sandpipers showed bimodality in migration strategy. Most showed no migratory preparation, but ~30% fattened, molted into breeding plumage, and performed partial post-juvenal wing molt (PPW) during the pre-migratory period. The frequency of PPW is positively related to culmen length (as a proxy for eastern breeding birds with a shorter migration distance). To decompose survivorship between migrant and oversummering (resident) Semipalmated Sandpipers, I used a multi-state model with 5 years of data and found survivorship 8 percentage point higher for oversummering juveniles and 21 percentage points higher for oversummering adults compared to same aged migrant birds, as expected as compensation for the loss of a breeding opportunity. I estimated annual survivorship with an open robust multi state model using 7 years of mark-resighting data from several thousand shorebirds marked at Paracas. As predicted by some migration theories, both species had higher annual survival estimates than those obtained previously at non-breeding sites further north. Western Sandpiper juveniles also had substantially higher annual survival estimates than adults, in line with the predicted survivorship benefits needed to offset their delayed reproduction. I found that the size of the survival advantage in juvenile Semipalmated Sandpipers is migration distance dependent. Western, but not Semipalmated Sandpipers showed a negative relationship in survival with the ENSO warm phase, probably due to the former’s closer association with the Pacific migratory flyway. Finally, I corroborated that the size of the survival advantage is distance dependent. My results provide novel information on non-breeding shorebird survivorship and perspective on the interrelationships that drive avian life history strategies. I confirm that Paracas is also a site with high demographic value.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Ronald Ydenberg
Department: 
Science: Biological Sciences Department
Thesis type: 
(Thesis) Ph.D.

Weighted l1 minimization techniques for compressed sensing and their applications

Author: 
Date created: 
2020-04-17
Abstract: 

Compressed sensing (CS) provides effective techniques for the recovery of a sparse vector from a small number of measurements by finding a solution to an underdetermined linear system. In recent years, CS has attracted substantial attention in applied mathematics, computer science and electrical engineering, and it has the potential to improve many applications, such as medical imaging and function approximation. One standard technique for solving the CS problem is l1 minimization; however, the performance of l1 minimization might be limited for many practical applications. Hence, in the past few years, there are many investigations into how to modify the l1 minimization approach so that better performance can be achieved. One such approach is weighted l1 minimization. In this thesis, we extend the weighted l1 minimization technique, traditionally used to solve the standard CS problem, to other applications. First, we develop a variance-based joint sparse (VBJS) algorithm based on weighted l1 minimization to solve the multiple measurement vector (MMV) problem. Unlike the standard l2,1 minimization method for this problem, the VBJS method is easily parallelizable. Moreover, we observe through various numerical experiments that the VBJS method often uses fewer measurements to reach the same accuracy as the l2,1 minimization method. Second, we apply weighted l1 minimization to the high-dimensional function approximation problem, focusing on the case of gradient-augmented measurements. The high-dimensional function approximation problem has many applications, including uncertainty quantification (UQ), where it arises in the task of approximating a quantity of interest (QoI) of a parametric differential equation (DE). For a fixed amount of computational cost, we see in various examples that, with additional gradient information, better approximation results are often achieved compared to non-gradient augmented sampling. Theoretically, we prove that, with the same sample complexity as the case of function samples only, the gradient-augmented problem gives a better error bound in a stronger Sobolev norm as opposed to an L2 norm. Finally, we use the adjoint sensitivity analysis method to compute the gradient information. As we show, this method computes the gradient samples of the QoI of a parametric DE with around the same computational cost as computing the samples of the QoI itself. We apply this approach to several parametric DE problems, and numerically demonstrate the benefits of incorporating gradient information into the approximation procedure.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Ben Adcock
Department: 
Science: Department of Mathematics
Thesis type: 
(Thesis) Ph.D.

A bivariate longitudinal model for psychometric data

Author: 
Date created: 
2020-04-30
Abstract: 

Psychometric test data are useful for predicting a variety of important life outcomes and personality characteristics. The Cognitive Reflection Test (CRT) is a short, well-validated rationality test, designed to assess subjects' ability to override intuitively appealing but incorrect responses to a series of math- and logic-based questions. The CRT is predictive of many other cognitive abilities and tendencies, such as verbal intelligence, numeracy, and religiosity. Cognitive psychologists and psychometricians are concerned with whether subjects improve their scores on the test with repeated exposure, as this may threaten the test's predictive validity. This project uses the first publicly available longitudinal dataset derived from subjects who took the CRT multiple times over a predefined period. The dataset includes a multitude of predictors, including number of previous exposures to the test (our variable of primary interest). Also included are two response variables measured with each test exposure: CRT score and time taken to complete the CRT. These responses serve as a proxy for underlying latent variables, "rationality" and "reflectiveness", respectively. We propose methods to describe the relationship between the responses and selected predictors. Specifically, we employ a bivariate longitudinal model to account for the presumed dependence between our two responses. Our model also allows for subpopulations ("clusters") of individuals whose responses exhibit similar patterns. We estimate the parameters of our one- and two-cluster models via adaptive Gaussian quadrature. We also develop an Expectation-Maximization algorithm for estimating models with greater numbers of clusters. We use our fitted models to address a range of subject-specific questions in a formal way (building on earlier work relying on ad hoc methods). In particular, we find that test exposure has a greater estimated effect on test scores than previously reported and we find evidence of at least two subpopulations. Additionally, our work has generated numerous avenues for future investigation.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Rachel Altman
Department: 
Science: Department of Statistics and Actuarial Science
Thesis type: 
(Project) M.Sc.