Topics in Time Series Analysis. IV. Various Aspects of Parameter Changes in ARIMA Models.
Abstract
In this paper we derive criteria for the detection of parameter changes in ARIMA models. An alternative approach is investigated in which we attempt to continually estimate or update the parameters as each new observation becomes available. Furthermore, we discuss how a step change in the parameters can affect the autocorrelation function. (Autor)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1976
- Accession Number
- ADA026310
Entities
People
- George C. Tiao
- George E. P. Box
- Johannes Ledolter
Organizations
- University of Wisconsin–Madison