ABSTRACT
The models traditionally used for seasonal and cyclical time series are stationary short memory processes on the one hand, or nonstationary processes due to a deterministic component such as seasonal dummies or to a stochastic trend such as seasonal unit roots. This work is reviewed in Sec. 2
inordertoplaceSCLMinsomeperspective.ThemodelingofSCLMis describedinmoredetailinSec.3.Section4discussesseveralparametricand semiparametricmethodsofestimationinSCLMprocesses.Testsofseasonal integrationandcointegrationarereviewedinSec.5.Allthisworkassumes knowledgeofthelocationofthepoles/zerosin.f(>..),asisreasonableina seasonalsetting,butnotnecessaryinacyclicone.Section6describes approachesforestimatingwinparametricandsemiparametricSCLMprocesses.Section7concludesthechapterwithsomementionofextensionsand applications.
