Andrew Smith (University College Dublin)

will speak on

Generating Gamma Processes

Time: 12:00PM
Date: Mon 2nd March 2020
Location: Seminar Room SCN 1.25 [map]

Abstract: Stochastic processes with Gamma-distributed increments arise in financial applications ranging from insurance claims, operational risk losses and loan defaults. Questions to answer using Gamma processes are often analytically intractable and so Monte Carlo is the numerical method of choice.

The generation of high-frequency Gamma process scenarios is also numerically tricky, owing to a large number of observations very close to, but not exactly, zero, combined with probability functions in standard software packages which fail for small values of the shape parameter.

Andrew Smith explains how the issues can be overcome with several new numerical algorithms for Gamma process generation.

(This talk is part of the Working Group on Statistical Learning series.)

PDF notice

Return to all seminars