|
Activity Number:
|
180
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #308045 |
|
Title:
|
Parameter Estimation of Diffusion Process with Sequential Monte Carlo
|
|
Author(s):
|
Rong Chen*+
|
|
Companies:
|
University of Illinois at Chicago
|
|
Address:
|
Dept of Information and Decision Sciences , Chicago, IL, 60607,
|
|
Keywords:
|
MLE ; diffusion process ; importance sampling
|
|
Abstract:
|
In financial markets and other applications, continuous time diffusion processes are often observed at discrete time. For nonlinear processes, the likelihood function of the parameters is often much easier to evaluate with continuously observed paths of the process than with discretely observed paths. In this paper we propose to use a modified version of sequential Monte Carlo method to sample the continuous path based on discretely observed observations. The sampled continuous paths are then used to estimate the likelihood function surface of the parameters. An enhanced importance sampling scheme is used to efficiently estimate the entire likelihood surface with a small number of samples. Empirical study and real applications are presented.
|
