This talk has been automatically
captioned. You can remove these by pressing CC on the video toolbar.
Name: Linda Mhalla
Talk Title: Discrete dependent extremes
Abstract: With view towards the current COVID-19 global epidemic, the study of time series of disease incidence is becoming increasingly important. In particular, we are interested in the temporal dynamics of the extremes of the number of fatalities during a pandemic. We propose constructions for time-dependent discrete generalized Pareto random variables, and specifically a hierarchical model for high discrete threshold exceedances by embedding a latent process with Gamma margins for the success probability of a geometric distribution. Based on pairwise-likelihood estimation of these models for the COVID-19 data in the US, we discuss the extremal dependence in the state-wise death toll while accounting for the inherent marginal non-stationarity through covariates.
This
talk is an invited talk at EVA 2021.