10th September 2007 — Copula models

Ricardo will discuss some elementary topics on copula models. Copula models are a fairly generic framework to construct multivariate distributions from marginal distributions, as well as measures of dependence. We will cover some basic definitions and applications. The main reference is Chapter 2 of the following thesis:

Dimitris Nicoloutsopoulos, Parametric and Bayesian non-parametric estimation of copulas, Department of Statistical Science, University College London, 2005.

In this meeting, we will also decide on topics to be discussed for the Fall term.