I am a Ph.D. candidate in the Department of Mathematics and Statistics, working under the supervision of Dr. Paul McNicholas. My research focuses on developing unsupervised learning techniques, specifically model-based clustering for mixed-type data. During this fellowship, I plan to implement and develop clustering and classification methods for analyzing multiplexed tissue images in pathology. This fellowship project will be co-supervised by Dr. Jonathan Bramson from the Department of Pathology and Molecular Medicine and Dr. Paul McNicholas from the Department of Mathematics and Statistics.

The aim of this project is to develop a robust and computationally efficient framework for
the statistical analyses of genetic variants using quantile regression. The proposed framework
aims to achieve similar performances with respect to power and computational efficiency as
OLS models under the correct model specification, while outperforming OLS models under
true complex interactions with unknown/unmeasured variables. The project will involve
empirical data analysis in the discovery of novel genetic variants for Body Mass Index (BMI)
as a complex phenotype with gene-environment components. This project will utilize the
publically data base of genotypes and phenotypes (dbGAP) on the NIH servers, and aims
to publish and disseminate any knowledge as a result of this work.

How Prescription Drug Coverage affects Health Services Utilization among Immigrants in Canada using Linked Data

I am a PhD student in the Health Policy Program, working under the supervision of Dr. Lisa Schwartz. My thesis explores how health insurance coverage impacts health service utilization and outcomes of vulnerable populations, such as refugees. I completed my Honours BSc in Human Biology (Health and Disease) at the University of Toronto and MSc in Global Health at McMaster University. My MacDATA fellowship is being co-supervised by Dr. Emmanuel Guindon and Dr. Arthur Sweetman. This research will examine how supplementary drug insurance coverage affects health care utilization by immigrants in Canada, combining health research methods, economics, political science and global health disciplines. Given the current federal government’s interest to increase the intake of immigrants into the nation, and the steady rise of pharmacare options onto the agenda, provincially and nationally, this research aims to inform these key policy decisions by presenting the effects of health policies pertaining to the health and well-being of future Canadians.

Fatemeh Armanfard is a Master’s student at the Department of Electrical and Computer Engineering, co-supervised by Profs. Jim Reilly and John Connolly. The focus of her thesis is on developing supervised machine learning techniques for analyzing electroencephalogram (EEG) signals recorded from comatose patients.

I am a PhD student in the Department of Economics here at McMaster. I completed my MA in Economics at Carleton University and my BA in International Economics at Toronto Metropolitan University (formerly Ryerson University). My research interests primarily lie in the fields of Labour Economics, the Economics of Education and Public Economics. Currently, I am investigating the factors that underlie the dramatic lack of applications to undergraduate-level engineering, mathematics and computer science programs (EMCS) compared to other STEM programs among female applicants. Employing data on all applicants to Ontario undergraduate programs, I incorporate factors including high school courses, peer effects and student socio-economic characteristics into various binary outcome models to assess application likelihood. This work has further implications for broader research on the gender earnings gap, as an absence of women from certain STEM professions (namely those involving engineering) has been identified as a factor in exacerbating the earnings disparity between men and women.

Automated Brain Injury Assessment- Machine Learning EEG/ERP Analysis

Trained both as a computer scientist (B.Sc., Dalhousie University) and a neuroscientist (M.Sc., McMaster University), my PhD work in biomedical engineering targets the development of automated assessment tools for brain function. Specifically, I utilize machine learning techniques to analyze brain signals, recorded using electroencephalography (EEG), to assess brain injury and its effects on different aspects of cognition.

Physical activity, sedentary time, sleep duration and health outcomes in preschoolers:
exploring the relationships with compositional analyses

The objective of the proposed project is to establish the relationships between
movement behaviours (physical activity, sedentary time and sleep duration) and health
outcomes in preschoolers (i.e., less than 5 years old) using compositional analysis.
Compositional analysis uses data that are proportional of a finite whole (e.g., 24 hours) and
can be used when all parts or just some parts of the whole have been measured (1,2). This
method considers the effects of a behaviour as a proportion relative to the other behaviours
instead of an independent behaviour. This is a novel approach to analyzing movement
behaviour data and it aligns with Canada’s newest Physical Activity Guidelines for Children
and Youth (i.e., 5 years and older). In 2016, the Canadian 24-Hour Movement Guidelines for
Children and Youth were published and this new version focuses on the interplay between
sleep, physical activity and sedentary time, rather than focusing on these movement
behaviours independently.

Mohammad is a first-year PhD student at Dr. Sue Becker’s lab. He obtained his BSc in Physics from Amirkabir University of Technology (Tehran’s Polytechnic) and his MSc in Cognitive Science jointly from the University of Trento in Italy and Osnabrueck University in Germany. His primary research interests are computational modelling of learning and memory and the brain’s dynamics. Specifically, he is working on computational modelling of traumatic memories in patients with post-traumatic stress disorder (PTSD) using artificial neural networks. His MacDATA project involves analyzing functional magnetic resonance imaging (fMRI) data, previously obtained from PTSD patients and healthy individuals, to investigate whether the patterns of hippocampal connectivity are different between PTSD patients and healthy individuals. Hopefully, the result of his MacDATA project will lead to a better understanding of which neural circuits specifically drive memory retrieval and the direction of information between these circuits in PTSD patients. In his non-academic life, Mohammad is a photography enthusiast and loves taking shots in various genres, such as nature photography and street photography.

Machine Learning and Bayesian Inference in Finance

I am a Masters student at the School of Computational Science and Engineering, with a focus on Machine learning, finance, and economics. Currently, my supervisor is Matheus Grasselli; the professor of financial mathematics and my co-supervisor is Dr. Anand for this fellowship. My main research is focused on how to apply machine learning techniques into finance due to the specialty of the capital markets. In other words, how to grasp the complexity of machine learning applications into investments. For this project, my supervisors will guide me using the Bayesian method to calibrate the work for option pricing.

I am an MSc student at McMaster University, studying statistics. I also completed my BSc (Hons.) at McMaster. My research focuses on machine learning, specifically outlier detection in Gaussian mixture modelling. This project will use machine learning algorithms to investigate the relationship between brain structure and psychosocial functioning and the link between changes in sleep and circadian rhythms and risk of postpartum depression. This project will be co-supervised by Dr. Benicio Frey from the Department of Psychiatry and Behavioural Neurosciences and Dr. Paul McNicholas from the Department of Mathematics and Statistics.

I am a third year PhD student in Cognitive Science of Language Department. I am also a researcher at the Reading-Lab in The Centre for Advanced Research in Experimental and Applied Linguistics (ARiEAL). My interests include newly emerging words and word learning. My research focuses on the dynamics behind the spread of newly popular words and the question of how these words are learnt by adult speakers. I have been conducting eye-tracking, EEG, and computational linguistics studies to answer these questions.

Project: Twitter data has been a popular way to examine newly popular words. My project aims to answer two questions by looking at the United States and Canada geo-tagged tweets: (1) what makes a word popular? (2) how have been these words spreading?

Cassandra D’Amore is a PhD candidate, supervised by Dr. Marla Beauchamp, in the School of Rehabilitation Science. Her research interests focus on aging, and include physical activity, mobility, participation, and end-user involvement. Cassandra’s thesis is using the Canadian Longitudinal Study on Aging (CLSA) to describe usual physical activity levels and behaviours in middle-aged and older adults. Her MacDATA fellowship work, supervised by Dr. Marla Beauchamp and Dr. Paul McNicholas, will expand this project by adding two additional analyses under this purpose. The specific aims of Cassandra’s project include 1) identifying latent physical activity behaviour patterns among middle-aged and older Canadians; and 2) exploring the relative importance of health-related factors in predicting these behaviours. The CLSA is nationally representative cohort study of approximately 51,000 middle-aged and older Canadians and contains 100’s of health-related factors. In her first aim, Cassandra will identify latent physical activity behaviour patterns that include not only moderate- and vigorous- but light-intensity activities, creating a more complete picture of physical activity patterns that may better inform physical activity promotion efforts. She will then use machine learning to identify the health-related factors that appear to be driving the prediction of physical activity behaviour patterns. This work will expand the current literature on determinants of physical activity by providing novel information using machine learning to identify the relative importance of predictors of physical activity.

Michael De Coste is a Ph.D. candidate in the Department of Civil Engineering. His graduate research focuses on the analysis and prediction of extreme ice events on Canadian rivers, applying machine learning and statistical analysis techniques. For his MacDATA project, he is co-supervised by Dr. Ridha Khedri from the Department of Computing and Software and his Ph. D. supervisor Dr. Zoe Li from the Department of Civil Engineering. This project aims to develop a data-driven model to predict the timing of the spring breakup of river ice in Canada on a national scale. This model will allow the strongest predictors of these events to be identified well ahead of the breakup, allowing forecasts with significant lead times to be made. This will provide decision-making support for communities directly affected by spring breakup while also providing insight on the impacts of climate change on river ice, both on a regional and national scale.

Big Data Genomics Infectious Disease Biosurveillance

I am a PhD Candidate in the Department of Anthropology, and I study a disease which has had a long and turbulent history with humans: The Plague. My thesis explores how ancient pandemics of plague are connected to modern outbreaks, and what interplay of factors leads to its re-emergence and dispersal. To examine this connection between the past and present, I work with a novel form of evidence: ancient DNA. My work in the McMaster Ancient DNA Centre involves extracting DNA from archaeological plague victims to capture, sequence, and reconstruct the genome of this ancient and deadly bacterium. However, to interpret this ancient genetic evidence, I need access to an extensive comparative dataset of genomes to robustly test the relationships between my ancient samples and modern disease outbreaks. With recent advances in sequencing technology, comparative datasets are plentiful but remain largely inaccessible and underutilized due to the challenges inherent in working with Big Data. In response, my project in collaboration with the MacDATA Institute, is to develop software tools to assist in querying online repositories for infectious disease metadata, and to streamline the data transformation steps needed to acquire and integrate raw data into comparative genomic projects. By improving data accessibility in the realm of infectious disease biosurveillance, this project seeks to advance current models of disease emergence and spread by giving researchers the tools needed to acquire crucial comparative data, particularly from currently underrepresented areas of the globe.

Arman is a first-year master’s student working under Dr. Andrew G. McArthur’s supervision in the Department of Biochemistry and Biomedical Sciences. His master’s thesis aims to contextualize and better understand antimicrobial resistance genes in bacteria by mining biomedical literature using natural language processing techniques. For his MacData fellowship, he will be examining the titles, abstracts, and metadata of papers discussing antimicrobial resistance to identify biases in the field and give insight into areas needing further exploration. With this opportunity, his fellowship may facilitate interdisciplinary training across antimicrobial resistance, natural language processing, and data sciences to understand the gaps in antimicrobial resistance and guide future research.

I am a Ph.D. Students at the Department of Economics at McMaster University. I am interested in Health Economics, Economics of Education, and Applied Econometrics. I also hold MSc in Financial Economics. Currently, I am working on the application of the minimax regret criterion to inform decision-making using data from actual clinical trials with multiple treatments and outcomes. In particular, my work includes assessing the performance of minimax regret decision rules compared to the conventional statistical inference-based approaches such as hypothesis testing, which are not ideally suited for decision-making problems. Additionally, I am interested in optimal trial design using the minimax criterion. Therefore, the ultimate goal is to improve the ex-ante and ex-post decision-making process under uncertainty. I am working under the supervision of two eminent professors, namely, Professor Arthur Sweetman from the Department of Economics and Professor Jean-Eric Tarride from the Department of Health Research Methods, Evidence, and Impact (HEI).