This talk has been automatically captioned. You can remove these by pressing CC on the video toolbar.

Name: Lekha Patel

Talk Title: Statistical learning of extreme spatio-temporal events with an application to global terror attacks

Abstract: Extreme events with potential deadly outcomes, such as those organized by terror groups, are highly unpredictable in nature and an imminent threat to society. In particular, quantifying the likelihood of a terror attack occurring in an arbitrary space-time region and its relative societal risk, would facilitate strengthened measures ensuring our national security. This talk focuses on describing our statistical modeling efforts to do so and presenting results from our approach using the open-source Global Terrorism Database (GTD). We first present a semi-parametric self-exciting model of attacks whose inhomogeneous baseline intensity is written a function of typical covariates found in the region of interest. Its triggering intensity is succinctly modeled with a Gaussian Process prior distribution, able to flexibly capture intricate spatio-temporal dependencies between an arbitrary attack and previous terror events. By inferring parameters of this model, we highlight specific space-time areas in which attacks are likely to occur. Furthermore, by measuring the outcome of an attack in terms of the number of casualties it produces, we then incorporate an extreme valued probability distribution of casualties, able to determine high-risk attacks in these regions. Our framework will last be utilized to study historic terror attacks in central and southern Asia.

This talk is a contributed talk at EVA 2021.