Evaluate Synthetic Populations

Overview

In agent-based modeling, researchers typically create of synthetic populations based on empirical data sources. These populations are then used alongside other transmission modeling decisions to simulate the spread of disease.

However, agent-based models are notoriously difficult to fit to traditional data

• e.g. incidence time series - due to the relatively large number of parameters typically present in such models. But agent-based models can also be made to yield a much broader range of outputs than traditional models, meaning they can be compared to other empirical measures that might otherwise be ignored.

This group will evaluate a particular synthetic population model and use network analysis approaches to calculate non-traditional outcomes, such “travel time” for a pathogen between one part of the population and another, and how that varies with behavior changes.

Things to consider

• This project will use a relative simple transmission model, and instead focus on the contributions of other elements of the overall model.

• This group is recommended for participants interested in…

  • agent based modeling
  • network analyses
  • spatial analyses

Background

This project will focus on the synthetic population associated with the analyses in this scenario analysis using the model generally described in this overview.

That model is descended from prior work on vector-borne disease, where this kind of assessment identified several opportunities for improving the overall model by improving the synthetic population.