Publication:
On the cumulated periodogram goodness-of-fit test in ARMA models

Thumbnail Image
Identifiers
URI: https://hdl.handle.net/10016/10721
Files
ws9642.pdf (327.76 KB)
Publication date
1996-07
Defense date
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Google Scholar
Export
Research Projects
Organizational Units
Journal Issue
Abstract
The asymptotic distribution of the cumulated periodogram goodness-of-fit test statistic for ARMA models is obtained, and is shown to be different from the limiting distribution of the standard Kolmogorov-Smirnov test statistic for probability distributions. The implications of this anomaly for inference purposes are analyzed. A modified cumulated periodogram goodness-of-fit test statistic is suggested, and its properties are studied and compared with other goodness-of-fit criteria proposed in the literature.
Description
Keywords
Autocorrelation function, Ljung-Box statistic, Normalized spectral distribution function, Residuals
Bibliographic citation
Collections
DES - Working Papers. Statistics and Econometrics. WS