Μελέτη ακραίων παρατηρήσεων μέσω της μεθόδου υπερβάσεων κατωφλιού και εφαρμογές στον κίνδυνο αγοράς
Modelling extreme observations via peaks over threshold method with applications in market risk

View/ Open
Keywords
Ακραίες παρατηρήσεις ; Υπερβάσεις κατωφλιού ; Κίνδυνος αγοράς ; Αξία σε κίνδυνο ; Αναμενόμενο έλλειμα ; Γενικευμένη Παρέτο κατανομή ; Εκ των υστέρων έλεγχος ; Μετοχές ; Κρυπτονομίσματα ; Μέγιστες τιμές περιόδωνAbstract
This master thesisfocuses on Extreme Value Theory (EVT) and its appli-cation to finance. Initially, the basic methodologies are presented, such as the Block Maxima and the Peaks Over Threshold method. Their theoretical basis, the distributions used (Gumbel, Frechet, Weibull, GEV, GPD) and the parameter esti-mation techniques are analyzed.
Next, the most commonrisk measures, Value at Risk (VaR) and Expected Shortfall (ES), which are used to quantify investment risk, are presented. Refer-ence is also made to backtesting, in order to check the accuracy and reliability of the estimates usinghistorical data.
The aim of thisthesis is to empirically compare three methods:Block Maxima, POT and Variance-Covariance using stock prices. The analysis aims to determine which method is more effective for estimating extreme risks and to offer conclusionsregarding the suitability of each approach under certaincondi-tions.
Finally, the above areapplied to real data employingthe statistical pack-age R. We compare the above-mentioned methods and draw conclusions about their effectiveness for Amazon, Apple,Netflix stocks as well as Bitcoin prices.


