Online Program Home
  My Program

Abstract Details

Activity Number: 457 - Zika Is Here, and We Need Statistics
Type: Topic Contributed
Date/Time: Wednesday, August 2, 2017 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics in Epidemiology
Abstract #323927 View Presentation
Title: Calibration and Evaluation of Agent-Based Models for Disease Modeling
Author(s): Shannon Gallagher* and Sam Ventura and William F Eddy
Companies: Carnegie Mellon and Carnegie Mellon University and Carnegie Mellon University, Department of Statistics
Keywords: agent-based models ; reproduction number ; effective reproduction number ; number of agents

Simulations are often used for prediction and inference of phenomena for which controlled experiments are not available, either due to impracticality or ethical concerns. One type of simulation, agent-based models (ABMs), is used to simulate the evolution of autonomous objects (agents) within their environment over time. In infectious-disease epidemiology, researchers use ABMs to analyze how prevention strategies such as vaccination or school closings will effect the spread of infectious diseases over time. Despite their importance to many real-world decision making processes, the statistical properties of ABMs are largely unstudied. We introduce an ABM that simulates the spread of an infectious disease for any administrative or geographic region in the United States. We analyze important quantities such as sensitivity to model parameters including the reproduction number, the effective reproduction number, and the number of agents sampled from the region of interest. We estimate parameters of our ABM so as to best fit incidence data, using data from past infectious disease outbreaks. Additionally, we discuss how ABMs may be calibrated for prediction and inference.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

Copyright © American Statistical Association