|
Activity Number:
|
415
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #305629 |
|
Title:
|
Independent Component Analysis for Multivariate Financial Time Series
|
|
Author(s):
|
David S. Matteson*+ and Ruey S. Tsay
|
|
Companies:
|
Cornell University and The University of Chicago
|
|
Address:
|
282 Rhodes Hall, ORIE, Ithaca, NY, 14853,
|
|
Keywords:
|
Time Series ; Independent Component Analysis ; Multivariate Volatility ; Conditional Heteroscedasticity ; Generalized Method of Moments ; Principle Component Analysis
|
|
Abstract:
|
We consider simple transformations of mutually independent univariate models as a flexible framework for estimating high dimensional time-varying conditional covariance matrices for financial asset returns. The use of prior information to impose constraints to provide more parsimonious models is illustrated. Increasingly flexible dynamics are demonstrated by allowing stationary, time-varying transformations. Asymptotic results suggesting efficient estimation methods, as well as procedures for component selection, will be discussed.
|
