Research

Hanlong Chen†, Malcolm Binns, Xinyang Feng, Aya Mitani

Master’s thesis · Dalla Lana School of Public Health, University of Toronto · Supervisor: Dr. Aya Mitani

Landmark models with shared frailty and hierarchical Gaussian-process longitudinal prediction for tooth-level survival, evaluated using within- and between-cluster AUC and Brier scores in simulation and electronic dental records.

Thesis defended · manuscript under internal review

Aya Mitani†, Hanlong Chen, Yuning Wang, Christopher McCulloch

Dalla Lana School of Public Health, University of Toronto · Supervisor: Dr. Aya Mitani

Agreement of probing pocket depths between two closely conducted periodontal exams, modelled with cumulative link mixed models with multi-level random effects and summarized by integral-based agreement and kappa statistics with delta-method variance.

Under review

Hanlong Chen†, Bruce Liu, Aya Mitani, Karim Keshavjee, Aziz Guergachi, Felipe Cepeda, Camellia Zakaria

SensAI Labs, University of Toronto · CPCSSN · Supervisors: Dr. Camellia Zakaria and Dr. Aya Mitani

A framework combining multiple imputation, mixed-effects trajectory modelling, and consensus clustering to study how imputation choices affect longitudinal patient phenotypes and uncertainty in cluster assignment.

Manuscript under internal review

Hanlong Chen∗,†, Ian Zhang∗, Bruna Seixas-Lima, Howard Chertkow, …, Malcolm Binns

Rotman Research Institute, Baycrest · Supervisor: Dr. Malcolm Binns

Latent-variable models and subtype analysis to distinguish general cognitive severity from age-related cognitive profiles in memory-clinic patients with dementia.

Manuscript under internal review

Ian Zhang∗,†, Hanlong Chen∗, …, Malcolm Binns

Rotman Research Institute, Baycrest · Supervisor: Dr. Malcolm Binns

RAMA jointly selects the number of principal components and clusters while downweighting redundant validation metrics for clinical subtype discovery.

Manuscript in preparation

Department of Statistical Sciences, University of Toronto · Supervisor: Dr. Murari Singh

Jeffreys and GIG-based priors for the coefficient of variation, with Gibbs sampling in OpenBUGS and R.

Undergraduate reading course · completed

Guangzhou Center for Disease Control and Prevention

Surveillance data cleaning and analysis of vaccination impact for outbreak response.

Completed