Methodology

This work develops advanced statistical modeling techniques to build efficient and scalable computational frameworks for investigating decision-making dynamics in behavioral tasks and neuroimaging measures.

This work introduces statistical methodologies for integrating data from clinical trials and and real-world data, which may share a common treatment arm but differ in their alternative treatment options, to learn robust and reliable individualized treatment rules.

This project develops a range of statistically principled methods for learning from missing, censored, and error-prone data, including boosting prediction, variable selection, joint longitudinal and time-to-event modeling.