|
Activity Number:
|
460
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #308874 |
|
Title:
|
Ensemble Sampling
|
|
Author(s):
|
Hedibert Lopes*+ and Nicholas Polson
|
|
Companies:
|
The University of Chicago Graduate School of Business and The University of Chicago Graduate School of Business
|
|
Address:
|
5807 South Woodlawn Avenue, Chicago, IL, 60637,
|
|
Keywords:
|
State-Space Models ; Stochastic Volatility ; Multivariate ; Markov Chain Monte Carlo ; Kalman filter ; Sequential Monte Carlo
|
|
Abstract:
|
In this project we provide a robust ensemble filtering and learning algorithm. Ensemble filtering methods are based on an update rule for particle propagation rather than resampling methods. Hence, the methods scale to high dimensions and avoid particle degeneracy. In the case of pure state filtering we compare our filter in a univariate stochastic volatility model with an the existing methodologies. We also provide a multivariate application to factor multivariate stochastic volatility.
|
