QRM 8-1: EVT meets TS (maxima of non i.i.d. observations) скачать в хорошем качестве

QRM 8-1: EVT meets TS (maxima of non i.i.d. observations)

Welcome to Quantitative Risk Management (QRM). In Lesson 8 we consider some final topics of TS. In this first video, we start by discussing what happens to the modelling of maxima, using EVT, when the underlying data are not i.i.d., but they rather constitute a proper time series (say an ARMA). We will see that some extra conditions need to be imposed, but that we can also characterise the dependence in the data in a brand new and useful way, using the so-called extremal index. Topics: 00:00 Welcome 01:14 Maxima of non i.i.d. observations 02:19 The D conditions 13:07 A more general Fisher-Tippett theorem 14:02 The extremal index 18:24 The block method