(1) Compile, contrast and develop data fitting techniques to address common issue of incomplete and imperfect covariate information when fitting many types of mechanistic models to data;
(2) Use a joint model framework to analyze the underlying correlation between key time series processes (such as daily number of cases, hospitalizations and deaths) and compare waves in a way that can give insight into how deaths and hospitalizations are changing in light of variants and vaccinations;
(3) Develop change-point models to measure the effectiveness of public health interventions that changes over time according to intervention timelines;
(4) Investigate the potential for using deep learning and ensemble classifier methods for classification of capacity utilization exceedance, when using Particle Markov Chain Monte Carlo methods to support joint estimation not only of model states over time, but also of parameter values.
Leads: Charmaine Dean (University of Waterloo, Waterloo) and Nathaniel Osgood University of Saskatchewan, Saskatoon)
Team members: Dongmei Chen, Joan Hu, Jeanette Jansen, Theodore Kolokolnikov, Juxin Liu, Felicia Magpantay, Bouchra Nasri, Chris Soteros