|
Activity Number:
|
294
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #300149 |
|
Title:
|
Inference for Long Memory Time Series with Application to Weather Derivatives Pricing
|
|
Author(s):
|
Nalini Ravishanker*+ and Jeffrey Pai
|
|
Companies:
|
University of Connecticut and University of Manitiba
|
|
Address:
|
215, Glenbrook Road, Storrs, CT, 06269,
|
|
Keywords:
|
Long-range dependence ; Temperatures data ; Vector ARFIMA models ; Weather derivatives
|
|
Abstract:
|
This talk describes inference for multivariate time series exhibiting long-range and short-range dependence, as well as heavy-tailed behavior, specifically for VARFIMA processes with underlying innovations assumed to be Gaussian, or Student-t or sub-Gaussian symmetric stable. Exact maximum likelihood estimation via the expectation-maximization (EM) or Monte Carlo EM (MCEM) algorithm, and fully Bayesian estimation via Markov chain Monte Carlo (MCMC) algorithms are described and compared to conditional MLEs. An application to pricing financial derivatives related to weather is discussed. This involves as a first step the modeling of multivariate daily temperatures at selected measurement sites in the US using the vector long memory models, paying attention to accommodating some degree of volatility exhibited by such series.
|
