Tensor Factorization Methods for Time-Varying Mediation Analysis in Continuous Glucose Monitoring Data
Developing tensor factorization methods for decomposing longitudinal continuous glucose monitoring data to study time-varying mediation effects.
Autoencoder-Based Causal Mediation Analysis for High-Dimensional Time Series: Application to Postprandial Glycemic Response in Type 1 Diabetes
Developing autoencoder-based methods to understand how meal carbohydrate intake affects post-meal glucose trajectories through insulin bolusing behavior in Type 1 Diabetes …