April CIDMATH/EAVE Seminar
You’re invited to join the April CIDMATH/EAVE Seminar Wednesday, April 1, 12 – 1 pm in CNR 1000.
This event will feature a presentation by Dr. Amy K. Winter, Assistant Professor at the UGA College of Public Health Department of Epidemiology and Biostatistics.
Title: Country-Level Predictability of Measles Outbreaks Worldwide
The introduction of safe highly effective measles-containing vaccines has led to large reductions in measles cases and deaths while also increasing the interannual variability in measles incidence. Fortunately, complex dynamical transitions of measles transmission may often have simple underlying exogenous explanations such as changing birth rates and vaccination rates. In this Gates Foundation–funded project, we sought to develop a global, data-driven measles outbreak prediction framework using supervised classification methods. We combined demographic, vaccination, and historical surveillance data to build and evaluate machine learning models, benchmarking against a baseline approach across diverse transmission settings. While our machine learning models did not consistently outperform the baseline, the systematic comparison of methods provided important insights. In this talk, I will describe the model development and evaluation process, present key comparative results, and discuss what these findings imply for measles control programs. More broadly, I hope to highlight lessons about the practical opportunities and limitations of applying machine learning to global infectious disease outbreak prediction.