Abstract #300532


The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.


Back to main JSM 2002 Program page



JSM 2002 Abstract #300532
Activity Number: 286
Type: Contributed
Date/Time: Wednesday, August 14, 2002 : 8:30 AM to 10:20 AM
Sponsor: IMS
Abstract - #300532
Title: Recursive Estimation of Misspecified MA(1) Models: Examples of convergence
Author(s): James Cantor*+ and David Findley
Affiliation(s): SAIC and U.S. Census Bureau
Address: 1100 North Glebe Road, Arlington, Virginia, 22201, U.S.A.
Keywords: Time series models ; pseudo-linear regression ; recursive maximum likelihood ; misconvergence
Abstract:

Stimulated by ideas of Hannan (1980), we have proved convergence results for misspecified MA(1) models for pseudo-linear regression (PLR--aka AML and RML1) and RML2 under monitoring. RML2 requires monitoring to ensure invertibility of the estimates. We present conditions that guarantee convergence under a simple, implementable monitoring scheme. Results are provided for MA (1) fit to AR (1) and MA (1) fit to MA (k), with additional analysis of MA (1) fit to MA (2) and a special case of MA (1) fit to MA (4), considered in a nonrecursive context by Tanaka and Huzii (1992). For these examples, we show that PLR converges, but to a value that is suboptimal (under mean square prediction error), whereas RML2 converges to the optimal value. These seem to be the first rigorous recursive estimation results for misspecified MA models.


  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2002 program

JSM 2002

For information, contact meetings@amstat.org or phone (703) 684-1221.

If you have questions about the Continuing Education program, please contact the Education Department.

Revised March 2002