Thursday, October 6, 2016 - 3:00pm
209 W. Eighteenth Ave. (EA), Room 170
Incorporating Hierarchical Structure into Dynamic Systems via Auxiliary Data
Le Bao, Pennsylvania State University
Dynamic systems have been used in a wide range of applications including molecular biology, weather forecasting, demography, etc. New technology advances and data collection tools allow people studying more localized dynamic systems. However, two technical barriers limit our ability to make reliable fine scale estimates: 1. the data availability can be imbalanced across localized dynamic systems leading to a high uncertainty in data sparse system; 2. the dynamic systems often do not have close form solutions and hence the parameter estimation is time consuming.
We propose a simple and innovative way to incorporate the hierarchical structure into the dynamic systems by using auxiliary data with applications of estimating of HIV/AIDS epidemics. It improves the predictive ability of the sub-national and sub-population disease dynamics especially in areas and risk groups with sparse data. It is implemented in UNAIDS recommended software that has been used by 163 countries in 2015 for estimation and projection of HIV epidemics.