TEST OF PERSISTENCE CHANGE UNDER THE INFLUENCE OF VARIANCE SHIFT

Jiawei Sun, Hao Jin
2017 International Journal of Advanced Research  
Int. J. Adv. Res. 5(11), 795-807 797 the observations implies a change in one of the errors and the converse is also true. Thus the test for a variance change can be performed based on the errors rather than the observations themselves. Furthermore, the test based on the errors outperforms the one based on observations since the latter is subject to serious power losses when the data is highly correlated. Thus, if the time series under investigation is stationary and invertible(see Brockwell
more » ... le(see Brockwell and Davis(1991), for the definition), then the former is naturally preferred (cf. Park et al.(2000)). In fact, the ease of application of the CUSUM of squares test lies in the fact that the limiting distribution of the test statistic is the sup of a standard Brownian bridge. For more historical accounts of change-of-variance analysis, we refer to Gombay, Horvath and Huskova(1996), Inclan and Tiao(1994) and the references therein.
doi:10.21474/ijar01/5833 fatcat:a4airqyvhjfrjmzllh5rj3jpc4