CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Hierarchical Bayesian agent based models and promise for causal inference Authors:  Samrachana Adhikari - NYU School of Medicine (United States) [presenting]
Abstract: While agent-based models are popular simulation models to assess hypothetical interventions and have been widely used in infectious disease modeling, HIV prevention studies and urban planning, among others, parameter inference and validation of such models remain a challenge. We will explore hidden Markov model representation of agent-based models, with a particular focus on infectious disease, that allows us to utilize Bayesian modeling and estimation tools such as particle filters for estimation and inference. Current methodological challenges and the potential of such a framework in assessing causality by simulating counterfactual will be discussed.