A solid background in probability theory and statistics is required, but no prerequisites in Extreme Value Theory. Illustrations on the R software will be made, but it is not requested to install R on your own computer. The tutorial will be organized into three sessions: Part 1: Introduction to Extreme Value Theory – Univariate Framework. Part 2: Estimation of tail parameters in the presence of random covariates. Part 3: Estimation of tail parameters in the presence of random censoring. This last part will be an introduction to the Plenary talk on the estimation of the conditional tail moment and reinsurance premium in the presence of censoring.
The tutorial will be organized into three sessions: (1) State space models and the Kalman filter, (2) Classical and Bayesian approaches to estimation for non-Gaussian state space models and (3) Sequential Monte Carlo, in particular particle filtering, for dynamic estimation. The textbook which related to (1) and (2) Time Series Analysis and Its Applications With R Examples (Shumway & Stoffer). The stochastic volatility model for financial returns will be the driving examples. R libraries for much of the code may be found at https://www.stat.pitt.edu/stoffer/tsa4/ .
Please follow the indications to access the vitual room: