Extreme Value Time Series: Modelling and Dependence

This lecture covers the motivation behind extremal limit theorems, the theory of point processes, and the applications of multivariate extremes. It discusses the modelling issues in extreme value time series, focusing on long-term trends, seasonality, and non-stationarity. The lecture explores the extremogram for stationary series and the implications of Theorem 37, highlighting the weak assumptions and the effect of local dependence on extreme values. It also delves into threshold sequences, Leadbetter's D(un) condition, and the extremal index, providing insights into the clustering of extremes and the mean size of clusters in a block.