Sparse prediction and anticipating the requests of declaration of natural disasters for a drought event in France
Người báo cáo: Nguyễn Thị Thanh Yên (Université Paris Descartes)

Thời gian: 14h Thứ 5, ngày 27/10/2022

Địa điểm: Phòng 507 nhà A6

Link online Zoom: 845 8621 8812

Passcode: 692956

Tóm tắt: Drought events are the second most expensive type of natural disasters within the French legal framework of the natural disasters compensation scheme. We develop a new methodology to anticipate which cities will request a declaration of natural disaster for a drought event, a key step of the national compensation scheme. The methodology hinges on optimal transport theory and an inertial proximal algorithm for nonconvex optimization. The optimization problem is designed so as to yield a sparse vector of predictions because it is known that relatively few cities will make the request.

Back