EVT and alternatives [11]

Go to: Summary | Previous | Next   
Bullet points include: Suppose interested in risk measures relating to losses, xj. EVT aims to supply two closely related results: Distributions of ‘block maxima’ (or ‘block minima’), i.e. maximum value of xi in blocks of m observations of x (more traditional use of EVT, wasteful of data): Distributions of ‘threshold exceedances’ (i.e. ‘peaks-over-thresholds’), where u is a predetermined high threshold and we focus on realisations of xj that exceed u (more relevant to, e.g. computation of Value at Risk i.e. VaR), i.e.:

Contents | Prev | Next | Library

Desktop view | Switch to Mobile