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

Name: Holger Rootzén

Talk Title: Real-time prediction of severe influenza epidemics using multivariate Generalized Pareto modelling

Abstract: Each year, seasonal influenza epidemics put high loads on health care systems and cause hundreds of thousands of deaths worldwide. A main concern for resource planning is the risk of exceptionally severe epidemics. This talk describes how multivariate Generalized Pareto models can be used for real-time prediction of the risk that an ongoing epidemic will be exceptionally severe, and for real-time detection of anomalous and potentially dangerous epidemics. The methods are applied to data from the French Sentinelles influenza surveillance system. The talk also describes a strategy based on standardized Brier scores, Precision-Recall curves and Average Precision scores for assessing the quality  of the predictions. Extreme Value Statistics has a large and rather unexploited potential for aiding healthcare and drug development, and the talk will also mention one more such opportunity:  evaluation of results from clinical trials.

This is joint work with Maud Thomas, Sorbonne University.

 This talk is an invited talk at EVA 2021.