You are here

Statistical Risk Metrics using Extreme Value Theory

Award No.: 
Principal Investigator: 
Hamid Mohtadi
PI Organization: 
University of Wisconsin-Milwaukee
This project consisted of three parts. First, we developed a new dataset on criminal/terrorist activities involving the use of chemical, biological or radionuclear (CBRN) agents on a global scale over the past 40 years. These include agents that would be the likely candidates to be used for any foodsector incident. Second, we constructed a risk metric using a novel statistical technique known as Extreme Value Theory to calculate probabilities and forecast the risk of such CBRN events at different casualty levels. Due to the large sample requirements of this method, however, we succeeded in our analysis only when we combined this dataset, in the third part, with a much larger dataset that we compiled, event-by-event, from an existing dataset (MIPT) of nearly 24000 observations.