A Simple Method Using CuScore to Monitor Changes in ARMA Coefficients

Author: Pan, Xia

Source: Communications in Statistics: Simulation and Computation, Volume 35, Number 2, April-June 2006 , pp. 515-530(16)

Publisher: Taylor and Francis Ltd

Buy & download fulltext article:

Price: $56.94 plus tax (Refund Policy)

Abstract:

CuScore was proposed by Box and Ramirez (1992) as a statistical process control scheme. Applying CuScore to ARMA coefficients has not been reported. This article proposes a simple yet effective method using CuScore to monitor coefficient shifts in ARMA process. The use is illustrated with examples of simulated and real production data. It can be seen that CuScore is effective in sequentially detecting ARMR coefficient shifts, hence in this term superior to the usual cumulative scheme such as CUSUM. A plan to distinguish coefficient shift and mean or variance shift in ARMA process is also recommended.

Keywords: ARMA coefficients; Control charts; CuScore; SPC; Time series

Document Type: Research article

DOI: http://dx.doi.org/10.1080/03610910600591388

Affiliations: 1: College of Business and Management, University of Illinois at Springfield, Springfield, Illinois, USA

Publication date: 2006-04-01