A0612
Title: Dynamic latent class modeling with applications
Authors: Wendy Lou - University of Toronto (Canada) [presenting]
Abstract: Motivated by a Canadian birth cohort study, where data were collected through multiple platforms across multiple sites over time, a robust approach is developed to describe various trajectories of subject changes (e.g. lung function) with associations of interest (e.g. gene, environment, etc). A class of dynamic models involving flexible mixture distributions and latent classes will be introduced, and practical examples will be given for time-varying measurements, such as nutrition and air quality. To illustrate the proposed approach, numerical results as well as real applications will be presented.
