Penalties For Misclassification Of First Order Bilinear And Linear Moving Average Time Series Processes

by Iheanyi S. Iwueze and Ohakwe, Johnson.

Abstract: Considering the similarity in the behaviour of the first order purely diagonal bilinear time series process (PDB(1)) and that of the linear moving average process of order one (MA(1)) under covariance analysis, the need arises for investigation of the penalty resulting from the misclassification of a bilinear process as a linear process and vice versa. It was discovered that the penalty function of misclassifying a PDB(1) process as an MA(1) process is non-negative while that of misclassifying an MA(1) process as a PDB(1) process is always negative. The penalty function of misclassifying a PDB(1) process as an MA(1) process is a polynomial of order 16 involving only even powers of the parameter of the PDB(1) process. On the other hand, the penalty function of misclassifying a MA(1) process as a PDB(1) process is a polynomial of order 2 or 3 with respect to the parameter of the MA(1) process.

Key Words: Diagonal bilinear time series model, moving average model, misclassification, penalty, polynomial function

Authors:
Johnson Ohakwe, johakwe@yahoo.com
Iheanyi S. Iwueze, isiwueze@yahoo.com

Editor: Peiris Shelton, shelton@maths.usyd.edu.au

READING THE ARTICLE: You can read the article in portable document (.pdf) format (173502 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 1796 times since JUNE 17, 2009.


Return to the InterStat Home Page.